STUDIES in BICYCLIC COMPOUNDS Thesis Presented to the University

STUDIES IN BICYCLIC COMPOUNDS

Thesis presented to the University of Glasgow

for the degree of Ph.D.

Richard Terence Wall, B.Sc.

1967 ProQuest Number: 11011830

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a com plete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest

ProQuest 11011830

Published by ProQuest LLC(2018). Copyright of the Dissertation is held by the Author.

ProQuest LLC. 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106- 1346 Myrett ACKNWLSDGEMEWTS

I should like to thank Dr, G.L. Buchanan and Professor

H.A. Raphael, F.R.S., for their constant encouragement and advice during the course of this research.

My thanks are also due to the staff of the departments responsible for recording high-resolution infra-red, proton magnetic resonance, and mass spectra, and micro-analyses.

The work described in this thesis was performed during tenure of a Research Scholarship awarded by the Carnegie Trust for the Universities of Scotland (1964-67)• SUI'MARY

Part I : Synthetic approaches to allohimachalol

2-(tf-f onnylbutyl )-2,6,6-1 rime thy 1 eye 1 ohe p t anone was synthesised in six steps from tetrahydroeucarvone. Attempts to effect aldol cyclisation to l^^jQ-tetromethylbicyclo (4«4»l) undec-7-en-ll-one, which has already been related to the naturally occurring sesquiterpene alcohol allohimachalol* were unsuccessful.

Other approaches afforded a variety of compounds, in particular, a number of bicycle (4»3»l) decanes derived from 1,5#5-trimethyl-

8-oxabicyclo (5*4«0) undec-6-en-9-one.

Part II : Studies in the 3-azabicyclo (3«3»l)nonane system

The reported procedure for the syntheses of a small number of 6-*substituted l,5-dinitro-3-raethyl-3-a'Zabicyclo (3»3*l)non-6-enes was extended to provide a wide range of 6- and 7-substituted analogues, the structures of which have been rigorously proved by chemical and spectroscopic methods.

Initial difficulty in assigning certain M R signals of these compounds led to a study of the NMR shifts of protons cc and JB to the nitrogen atom in simple amines and amine picrates. Some correlation between structure and magnitude of shift was observed.

The scope of the reaction sequence leading to these compounds was thoroughly investigated, and a. number of 3-un substituted, 3-ethyl, and 3-benzyl analogues prepared. In particular, this made available several 3-substituted l,5-dinitro-3-azabicyclo (3.3»l)nonan-7-ones,

IR studies of which established the absence of interaction between the amine and carbonyl functions.

Evidence was presented confirming reported findings of the chair-chair conformation of the 3-azabicyclo (3*3»l)nonane system. CONTENTS

PART I t Page

Introduction ...... 1

Formulae

Discussion ...... 8

Tables

Formulae

Experimental ...... 29

References ...... 58

PART II •

Introduction ...... 62

Formulae

Discussion ...... 69

Tables

Formulae

Experimental...... ••••••• 104

References ...... 145 PART I

SYNTHETIC APPROACHES TO ALLCHIMACHALOL -1-

IMTRODITCTIQN

Organic chemistry has gained much from the study of bridged bicyclic molecules, but in the main, attention has been focussed on the naturally occurring bicyclo (2.2.1) heptane and bicyclo (3 .1 .1) heptane systems; larger ring systems have received less attention, owing to a scarcity of synthetic methods. In particular, there is difficulty in constructing a bridge of more than three carbon atoms, e.g. in the preparation of a bicyclo (4 «n.l) alkane from a (n+3)-membered cycloalkanone (l). Hence although closely related in structure, bicyclo (4 *3*1) decanes (2) and bicyclo (4*4 *1) undecanes (3) differ markedly in frequency of occurrence in the literature.

A survey of the methods available for preparations of bicyclo

(4.3.1) decanes and bicyclo (4 *4 *l) undecanes reveals that there are four general routes, designated A,B,C and I.

Method A involves ring expansion of a suitable fused bicyclic system, of which there are three versions (a)—(c), as shown in Scheme 1.

The following examples illustrate its application to the syntheses of bicyclo (4*3*1) decanes and bicyclo(4*4*l) undecanes.

Vogel et al.1 have reported the synthesis of (4)> which behaves as the bicyclo (4 #4 *l) unde cane aromatic compound

1 ,6-methanodecapentaene1 (4a)* This is an example of the noccaradiene- cycloheptatciene equilibrium. Also known is its negatively charged bicyclo (4 .3.I) analogue, the decatrienyl anion (5)* Analogues bf -2-

(4 ) occur in the steroid field. Treatment of 19-hydroxy^ androst-4-,ene-3»17-dione (6) with the amine (7) afforded^ the two

5,19-cyolosteroids (8) and (9)* The latter was converted into the bicyclo (4*4*1) undecane steroid (10).

Likewise, Dauben and Laug^ observed that treatment of

Jbrans-8-hydrindanylcarbinylamine (11) with nitrous acid produced a mixture containing 6% bicyclo (4*3*l) decan-l-ol(l2). Similar results were obtained with trans-9-decalylcarbinylamine (13) • 4 5 Although in this case details were not given, ’ one product would be bicyclo (4«4*l) undecan-l-ol (14)t probably in the cis and trans forms, since Eliel^ attributed both to Dauben and Westman (via a private communication).

A novel skeletal rearrangement in the steroid field has recently been observed.^ ^3-bromo-/p , 5-epoxy-5^- cholestan-3p-ol(l5) when treated under reflux with lithium aluminium hydride in tetrahydrofuran formed 92% of 4r 5-seco-4y6-cyclo—^J-cholestan-

3p,5*-diol (16), a 6 ,10-cis-bicyclo (4.3.1) derivative with ring-® constrained in the boat conformation.

In method B, a cyclodecane or cyclononane compound undergoes 0 a transannular reaction (Scheme 2) • Westman and Ssevens obtained from the Bolvolysis of 6-methylene-cyclodecyl tosylate (17) a mixture of 93% cis-bicyclo (4.4.1) undecan-l-ol (14), and 7% of an olefin fraction, the Tna.-in component of which, had a 6 LC retention time identical with bicyclo (4*4*1) undec-l-ene(l8), Comparison was obviously available with work done by Dauben, Westman and Bond,^ which has not been fully described#

Several transannular reactions of this type were observed by Buchanan et al.^ (vide infra), e.g*, purification of the cyclononene gem-diester (19) by chromatography caused partial conversion to the bicyclo (4#3*l) ketoester (20).

The remaining two methods, C and L, may be considered together# They have been applied to the syntheses of bicyclo-(4*3*l) decanes, but not to the syntheses of bicyclo (4 #4*1) undecanes#

Basically, they consist either of attachment of a three carbon chain to a cycloheptane ring and subsequent ring closure (method C), or the opposite process, viz#, condensation of a four carbon unit to a cyclohexane ring (method D)# In two examples of D, bis-1,3- condensation to the cyclohexane ring yielded the bicyclo (4*3*1) decane in one step#

The first example of method C was reported by Prelog et al,10 in 1949# Condensation of 2-carbethoxycycloheptanaifiwil4i l,3~dichlorobut-

2-ene afforded 1-carbe thoxy-1-( 3-chlorobut-2-enyl) cycloheptanone (21), which was cyclised with concentrated sulphuric acid (with prior formation of the 3“hetobutyl derivative, a reaction discovered by

Wichterle11) to 1-carbethoxy-7-nethylbicyclo(4*3*1) dec-7-en-10-one (22),

Hydrolysis and decarboxylation led to I^cthylbicyclo (4*3*1) dec—7~en—10-One (23)* Removal of "the oaa.'bothoxy giroup from (2l) was found to cause cyclisation in the alternative way to form theanone (24). 12 A method discovered in 1956 by Stork and Landesnan for preparing bicyclic aminoketones by condensation of acrolein with, the pyrrolidine enamines of cy cl open t anon € and cyclohexanone was used by Grit ter et al.**^ to prepare 7-py^rolidinobicyclo (4 .3*1) decan—10-one (25) from the pyrrolidine enamine of cycloheptanone.

Later investigations on this reaction by Schut and LiuM included the preparation of the related 7-(4-phenyl-l~j»iperazyl) bicyclo (4 .5 .I) decan-10-one (25a),

In 1963, Sands^, using the Wichterle reaction, synthesised bicyclo (4 .3 .I) dec-7-en-10-one (26) from 2-carbethoxy cycloheptanone, or, more economically, from cycloheptanone, and 1,3-dichloropropene, followed by cyclisation of the resultant 2-(3-chloroallyl)- cycloheptanona (27, 27a) with concentrated sulphuric acid. The use of

1,3-dichloropropene instead of l,3-dichlorobut-2-sne (vide infra) for the preparation of bicyclic compounds offers the same advantages of an allyl halogen for the alkylation of the carbethoxycycloalkanone, together with a vinyl halogen, inert in the initial alkylation, but available for the later sulphuric acid-induced cyclisation. Furthermore,

2-(3-chloroaliyl) cycloheptanone does not have a methyl group available for reaction with the carbonyl group when the carbethoxy group (as in 27a) is absent, and the alternative mechanism of ring closure

is therefore not available.

The same author^ prepared 1-carboxybicyclo (4* 3*1)

dec-7-en-10-one (28) from 2-carbethoxy cycloheptanone by means of

the method used by Cope and Synerholm^ to synthesise the analogous

bicyclo (3.3.1) compound from acrolein and 2-carbethoxycyclohexanone.

Hydrogenation converted (28) to 1-carboxybicycl© (4 *3*1) decan—

10-one (29).

Luring their investigations on reactions involving bridge

scission of bridged ring systems, Buchanan et al.^ condensed acrolein

with 2-cartetkcxy-7-ne thylcyclohep tanone (30) to obtain the bicyclic

alcohol (31), but all attempts to dehydrate this to (20) yielded the

tetralin acid (32) as the only recognisable product.

Thpi preliminary stages of a stereoselective hydroazulene

synthesis by ilarshall and Partridge^® involved condensation df

2—carbetfacaQrcyclcheptanone with methyl vinyl ketone, and ring closure

with suljgtouric acid to yield (.22), the same compound prepared by

P r e l o g et al.10. KLaboratian to the saturated alcohol (33), followed

by solvolysis of the mesylate (33a), afforded Hie hydroazulene (34)

in QOf yield*

"TTTbii^ bridged ketol (35) was obtained in Bp- yield as a side—

product free Hhe cyclisaiioii of the trione (36), (prepared by

The first example of route D to bicyclo (4*3*l) decanes 20 was due to Opitz and Milderiberger , who in 1961 prepared the

ketones (38; 15%) and (39; 31%) by refluxing 1,4-diiodobutane

and o—xylylene bromide respectively with the pyrrolidine enamine of

cyclohexanone and ethyldicyclohexylamine in acetonitrile. The

ketone (38) was very slightly contaminated with (40), arid another

product in this reaction was the dione (41)* However, similar

treatment of cycloheptanone enamine yielded no bicyclo (4.4*1) undecanones, only starting material being obtained. 16 Sands reacted 2-carbethoxycyclohexanone with 1,4“

dichlorobut-2-ene to obtain crude (42) in low yield. Hydrolysis and

hydrogenation furnished the ketoacid (29), which he obtained by an

independent synthesis (vide supra)•

As proof of the aforementioned steroid rearrangement, n Collins et al. treated the tosylate (43), prepared from cholest-

4-en-3-one, with potassium t-butoxide to form in low yield the ketone

(44), the main product being the cyclic enol ether (45). Another

recent synthesis, due to Loewenthal2!, involved the cyclisation of the

cyclohexenone acid (46) by refluxing with 2-naphthalene sulphonic acid

in toluene to the bicyclic enedione (47) in only 8% yield, althou^i analogous bicyclo (3*2.1) octanes and bicyclo (3*3.1) nonanes were

formed in 80-91% and 70% yields respectively. -7-

It can be seen from the above survey that there is no preparative route to bicyclo (4.4.1) undecanes.

No bicyclo (4 .3*1) decane has been isolated from natural

sources. In contrast however, Bisarya and Sukh Dev^, investigat

ing the alcohol fraction (3.5$) of the essential oil from Himalaya

deodar (Cedrus deodara, Loud.) found it to contain two new

sesquiterpene alcohols, himachalol (48) (41$) and allohimachalol

(30$), the latter possessing the bridged bicyclo (4 .4 .1) undecane

structure (49)> a new type in sesquiterpenoids. It was postulated 2^ 23 to have arisen from the species (50) , the biogenetic precursor of himachalenes, Icngibomeol and related compounds. This was confirmed

by solvolysis of allohimachalol tosylate to yield ot-himachalene

(51* 3$) > P -himachalene (52; 15$) > himachalol (48; 24$) and

allohimachalol (34$) > (Scheme 3). The only possible structures for

allohimachalol were hence (49)>(53) and (54). Spectroscopic and

chemical evidence ruled out (53) and (54)* The solvolysis experiment

also permitted the unequivocal assignment of the configuration of the hydroxyl group, and since the absolute configuration of the himachalenes 24 and himachalol have been established , allohimachalol must have the

absolute stereostructure (49).

Because of the novel structure Of allohimachalol (49)> the

absence in the literature of a planned synthetic route to a bicyclo (4*4*1)

undecane, and the interest in this department in bridged bicyclic systems2^,

a synthetic sequence from carvone (55) aimed at allohimachalol was undertaken* CD-CD 4a 4a

SCHEME 1 c (°) (b)

(b)

a > c Et^MCFCHCIF

R = 0M eJ0A cJ-f O CH2NH2 OH

HoA " < ^ r H r*v

13 I O i / 4 ^ - HO'

O T s

17 18

SCHEME 2

19 20

22 23 2 U n v r\ ^ c r > : U c: .; j r \

25 R = - 26 27 R = C 0 2B c

25a R N N-Cp 27a R = H

c o 2h

31 32 33 R = H 33a R = M s

34 3b J 0 j w 1 0 VJ

38 39

(CHo).

42 43 44

4 5 48 SCHEME -8-

DISCUSSIQR

The readily available monoterpene carvone (2 ) was selected as the starting material for ci, synthetic route to allohimachalol (1), the only known naturally occurring bicyclo

(4 .4 .I) undecane. The long-known conversion of carvone (2) to eucarvona (3) could then be utilised, followed by hydrogenation to the saturated • cycloheptanone, tetrnhydroeucarvone (4), (Scheme A).

The sequence then required would be essentially the attachment of a four-carbon bridge to this ketone in the 2*7~positions (Scheme B).

It was anticipated that this would be a difficult task, for all syntheses ox the analogous bicyclo (4 .3*1) decane sysGem bjr elaboration of a cyciohexanore have proceeded in Icvr yield (see introduction).

It was therefore decided to use as an intermediate the

1,5-dione (10), the proposed sequence to which involved the facile attachment of a three-carbon side chain to the 2-position of tetrahydroeucarvone (Scheme C). Condensation of (4) with acrylonitrile and hydrolysis ox'iha resultant ketouitrile (5) would afford the ketoacid

(6), cyclisation and concomitant dehydration of which would then yield the &-enol lactone (7). By analogy with other enol lactones, it was thought that (7) would react with a methyl Grignard reagent to form the bicyclic ketol (9), convertible to the desired dione (lO) by a retro- aldol process. Since the side chain carbonyl group of (10) would be less hindered than the ring carbonyl, treatment with the Wittig reagent methoxymethylenetriphenylphosphorane would furnish the enol -9-

ether (25) of the ketoaldehyde (48)« By cyclisation of the side chain of (48), the required bicydic skeleton would be obtained, in the form of the enone (50), (Scheme D) , which was prepared by 22 Bisarya and Sukh Dev by oxidation of allohimachalol.

An alternative route to the key intermediate (4 8 ) envisaged elaboration of the enone (L3) t 311 aldol product of (10), Treatment of the enone (13) with me thylene tr iphenylpho spin or ane would produce the diene (hi). It was hoped that advantage could be taken of the higher reactivity of the exocyclic double bond to reduce (41) selectively to the olefin (44) ? oxidative cleavage of which, by ozonolysis or by osmium tetroxide and sodium, periodate, would then yield (48).(&h«i*« 6 ) . 26 Carvone (2) was brominated and debydrobrominated to yield 27 eucarvone (3) by the method of Wallach. Hydrogenation to 2 3 tetrahydroeucarvone (4)^7 "the hethcd cf Barnes and Houlihan proceeded in a higher yield than tha,t reported. Tetrahydroeucarvone, a colourless, volatile oil, with a strong characteristic odCur* had v

^ q 4 1704 cm. ^ A feature of its M R spectrum useful for structural confirmation of certain compounds derived from it was an AB quartet at 7*551* (20, J = 11 • 5 c/s) due to the two non-equivalent C7 protons. 29 ' Under strongly basic conditions, tetrahydroeucarvone underwent smooth cyanoethylation to afford 57f° of a single product, which when purified was obtained as a colourless crystalline solid,

Cl5H2iON, m.p. 44*5 “ 46°. Its mass spectrum confirmed the molecular weight of 207, i.e. monocyonoethylation had occurred to form (5 ).

The XR spectrum had peaks at 1703 (C=0) and 2255 cm. ^ (c = N), and in the HMR spectrum, an AB quartet at 7*55 X (2&» J=ll»5 c/s) demonstrated the absence of substitution at C7.

Hydrolysis of the ketonitrile (5) by prolonged reflux with dilute hydrochloric acid in dioxan proceeded in almost quantitative yield to the ketoacid (6 ), a colourless solid, C13H2203* I^s structure was rigorously confirmed spectroscopically as for (5).

Treatment of (6) with refluxing acetic anhydride, in the 3°,23 presence of sodium acetate afforded the &-enol lactone (7), a distilled sample of which, a colourless vd.scous liquid, exhibited -11-

absorption at 1758 (enol lactone 0*0) and 1662 om."’1 (0 s C) in the IR. A sharp singlet in the M R spectrum at 4*?3 Z was assigned to the olefinic proton* The material appeared to be hygroscopic, and a satisfactory analysis was not obtained. The mass spectrum however showed a molecular weight of 208.

The enol lactone (7) remained unchanged on exposure to 31 a 0*5 M excess of methyl magnesium iodide for one hour. When 32 subjected to a 5 M excess for three hours, at either room or reflux temperature, however, (7) was converted completely into a mixture of three compounds. Chromatographic separation afforded the most polar, predominant component, as a viscous colourless oil after distillation. M R and mass spectroscopy showed it to be the ketol (8), and not the desired bicyclo (4.3*1) decanolone (9).

The material had IR absorption at 1696 and 3&10 cm. , and weak, broad, concentration dependent absorption at 3530 and 3370 cm.”l, assigned to an intermolecularly bonded hydroxyl. In the M R spectrum, a quartet at 7*58'® J = 10*5 Q./&, O7 protons) excluded the bicyclic structure (9); also observed were a singlet at 8*32 X

(hydroxyl proton), removed on adding D2O» acd a singlet at 8*80 X

(6h, side chain methyl protons), as well as the signals due to the

C2- and C6-methyl protons. The mass spectrum had no molecular ion, but contained strong peaks at 222 (P-18, loss of H20) and 154 (P-86, loss of side chain) ^/e* This latter peak was found to be common to -12-

all compounds of this type, arising by loss of the side chain

(table5 )* These data fully confirmed structure (8)»

That the desired 1,5- diketone (lO) was an intermediate in this reaction was proved by its isolation from the above chromatogram in very low yield. Physical data were assembled later, when more material became available ( vide infra).

Since the preparation of (LO) fhom (7 ) was clearly impractic able, attempts were then made to find an alternative route from the ketonitrile (5) and the ketoacid (6)* Considerable starting material / /— \ 33 was recovered by treatment of (3) with methyl magnesium bromide.

A complex mixture containing only a trace of unchanged (5) resulted 33 when methyl magnesium iodide was used. When reacted with methyl 33 lithium, the ketonitrile (5) yielded, on removal of acidic material, a mixture of three compounds. Separation by preparative TLC afforded the dione (lO) in 2$$ yield. Unfortunately, the course of this reaction was not reproducible, and later runs gave distinctly lower yields (ca. 7 of (lO). This route to (10) could not therefore be regarded as suitable for inclusion in a synthetic plan.

Methyl ketones may also be prepared by the action df methyl 34 t C\ lithium on acids. Accordingly, the ketoacid (o) was refluxed for two hours with a 3M excess of this reagent. The product consisted almost entirely of unchanged acid, the small neutral fraction being a mixture of three components.

Another method which was considered was the reaction between -13-

35 dime thylcadmium and an acid chloride* However, attempts to prepare the acid chloride (11 ) by reacting the acid (6) with either thionyl chloride or oxalyl chloride led to mixtures of (H) and the enol lactone (7), the latter predominating. An odour of acid chloride was detectable in the product, which had HR absorption at 1700

(cycloheptunone 0=0), 1800 (acid chloride 0*0) and 1760, 1660 cm. ^

(enol lactone absorptions), Also, the NMR spectrum was similar to that of pure (7 ),

Certain esters ane convertible to p-ketosulphoxides by reaction with two equivalents of the sodium salt of dimethyl sulphoxide (dimsylsodium). Hydrogenolysis results in the formation 36 of the corresponding methyl ketone. Treatment of the acid (6) with diazomethane gave in high yield the ester (12), which analysed for

C14H24Q3, had I698 (ketone) and 1742 cm,~^ (ester), and a NMR singlet at 6*33 X (3K, methyl ester protons). Only a low yield of neutral product, accompanied by much ketoacid (6), was obtained on reaction of (12) with dimsylsodium. It contained two compounds, and its IR spectrum was transparent in the carbonyl region,

A successful route to (IQ, practicable on a large scale, was finally achieved by condensation of tetrahydroeucarvone (4) and methyl 37 vinyl ketone, with ethanolic potassium hydroxide as catalyst. A

40i<> yield of (L0 ) was realised, after distillation to remove unchanged

(4), and chromatographic separation from a small quantity of the enone

(13), formed by cyclisation of Q 0 ) in the basic medium. The -14-

dike tone (lO) formed a mono-2,4-dinitrophenylhydrazone,

^20^28^4^5» which had a single carbonyl peak in the 3E at 1699 cm. ^

Since the present dione hadv^^4i695 (cycloheptanone 0=0) and

1721 cm, (sidechain C=0), the more hindered ring carbonyl group had therefore been unaffected by 2 ,4^-dinitrophenylhydrazine# Hie

side chain methyl group gave rise to a singlet in the NMR at

7*89 X (?^)» The molecular weight of 224 was confirmed by mass

spectrome try.

With a practical route to (10) available, its elaboration

to (48) could now be studied. However, simultaneously, it appeared

of interest to investigate the potential of the readily obtainable enol lactone (7). This was prompted by the discovery in this

departmeht that compounds of this type, e.g., (1 4 ), undergo reductive 38 rearrangement with lithium aluminium tri-( t-butoxy) hydride to form 25 bridged bicyclic ketols, e.g., (15)• Similar treatment of the enol lactone (7) in TEF at -70° afforded a mixture (in decreasing polarities) of the ketoacid (6), the bicyclic ketol (16), unchanged (7), and a small amount of a compound of unknown constitution. Chromatographic separation furnished 39% of the desired bicyclic ketol (16), obtained after distillation as a colourless crystalline solid, m.p. ca. 7I-760, which was not recrystallisable. Chromatography (TLC and GLC) indicated the presence of only one compound, later shown to be the axial alcohol

(vide infra). Some later runs of this reaction gave material with -15-

IR absorption identical to that of the axial ketol, but consisting of

two compounds, the axial and equatorial epimers. The mass spectrum of axial (16) showed a nolccular ion at P=210 m/e, with a peak at

192 m/e (P-18j loss of H2O). IH bands were observed at 1698,

3615 and 3445 cm."1, the last being the centre of broad, concentration dependent absorption, due to the intermolecularly hydrogen bonded hydroxyl group. The NMR spectrum showed a broad multiplet at

5*50 X (1H, half-band width ca. 9c/s, C7 carbinyl proton, i.e., 39 the C7 hydroxyl group must be axial ), a doublet at 7*78 X

(IH, J=3*5c/s, C£ proton), a singlet at 8*1 X (hydroxyl proton), which disappeared on addition of D2O, as well as the signals due to

the ring methylene protons, and the Cl and C5 methyl protons (Table7 ).

The ketol (16) was further characterised by the preparations

of the diol (L7) and the ketotosylate (18). The diol was available

in rather low yield by prolonged treatment of (16) with lithium

aluminium hydride or lithium aluminium tri-(t-butoxy) hydride, followed by chromatographic removal from unchanged (16), which was

always present. This reflected the lack of reactivity of the C10

carbonyl group, which is discussed below. The diol (17), C13H24O2, with in.p. 91-97° (C]_q epineric mixture), halV^^4 3625 (free hydroxyl) and 3550-3200 cm."1 (intermolecularly bonded hydroxyl),

The presence of two hydroxyl groups was shown by the NMR spectrum, in which the C7 and carbinyl protons absorbed as a triplet at -16-

6»32 X (J=5c/s) and a doublet at 6*45 X (J«2*5o/s) .respectively.

Treatment of the ketol (16) with tosyl chloride in pyridine produced the ketotosylate (18), m.p. 111-115°f which analysed for C20H28O4S, Spectral data were completely compatible with structure (18). In particular, the It'S spectrum contained a broad multiple! at 4*80 X (iH, half-band width ca, Sc/sj C7 carbinyl 39 proton, i.e.f the Gy tosyloxy group must be axial ) and the mass spectrum, although containing no molecular ion, showed a peak at

192 m/e (P--172, loss of p-toluenesulphonic acid).

The Cy — axial conformation of (l8)> and therefore (iq), was confirmed by the conversion of (18) by refluxing ethanolic sodium 4° * \ ethoxide to the urccujugated enone (19) in good yield. This compound, an extremely volatile oil with a camphoraceous odour, had a mass spectral molecular weight of 192. The vinylic protons gave rise to a jMR. multiplet at 4*13 Z (?IT, half-band width 5 c /3 ), showing 41 that, as expected by Bredt's rule, the double bond was unconjugated*

Also observed were two unresolved doublets at 7*45 Z (C£ proton) and

1*72 X (C9 protons, J ca. 2c/s). v were observed at 3030

(=C-H str#) and 1703 cad'1 (C-0), and at 764 and 694 cm.

(=C-II def.), The UV spectrum had \ fairly typical 70 absorption for a & —unsaturated Ketone.

Oxidation of the ketol (16) by chromium trioxide in sulphuric 42 acid gave the dione (20) in hi$i yield. A twin carbonyl band was -17-

observed in the IH spectrum (in CCI4), at 1722 and I696 cm."1

Although this doublet is explicable on the basis of eyoloh^xanone and cycloheptanone carbonyl frequencies respectively, compounds similar to (20), incorporating a non-enolisable cyclic diketone moiety are known to exhibit a twin IR carbonyl poak.^* ^ Of the two explanations of this phenomenon, Fermi resonance and vibrational coupling, the latter has been preferred. In the case of the dione

(20), the possibility that .Fermi resonance might be the cause of the doubler carbonyl was excluded by the observation that the peak separation (Av= 25 gei.""^) was virtually unchanged in a series of solvents of differing polarities (Table3). The dione formed a mono-2,4~dini fcrophonylhvdrazons , C19H24N4O5 > which hadV 1695 u—u due to the unreacted bridge carbonyl group, In the MIR spectrum, the

Cg bridgehead proton gave rise to a singlet at 7*00 XJ (iH), very slightly split, possibly by long range coupling with the (J4 protons or the Gcj me thy'] protons,. Complex absorption at 7' 2-7* 5 (2H) was attributed to the C'q protons0 .Hass spectrometry confirmed the molecular weigh I; of 203 «

Two attempts wore made to effects a shorter roirbe to

(20) from the readily avariabie ketoac*.d \b j ar;d. the ketoestea Cl2 ) •

A mixture of the acid (6 ; and p -•tolaenesulphcnrc acid in tetralin was -44 heated at reflux in a bean-Stark water separator. The product, although almost entirely non-acidic, consisted of a complex mixture. -18-

45 Marshall and Scanio prepared the bicyclo (5.5.1) undecane dione

(21) by cyclisation of the ketoestor (22) with sodium hydride in refluxing 1,2—dimethoxyethane. However, similar treatment of the ketoester (12) afforded mainly the acid (6), and none of the dione (20) was detectable in the neutral fraction.

Before a practicable route to (10) had been discovered, it was realised that the bicyclic dione (20) could be of use in the preparation of 0-0). The CpQ carbonyl group of (20), known to be unreactive, should remain unaffected by a methyl Grignard reagent, and the resultant ketol (9) convertible to (lo) by a retro-aldol process. In the event, exposure of (20) to excess methyl magnesium iodide furnished a mixture of unchanged (20), the two epimeric ketols

(9), and the diol (2 3) • The formation of this complex mixture made it apparent that the synthesis of (10) by this method was not feasible, and with the - IR and HMR spectral characterisation of the crystalline diol (23) completed, attention was turned to the behaviour of (20) under the following conditions. Reaction with methanolic base gave only the ketoacid (6 ). The alternative cyclononanone acid

(24), which would result by attack at the bridge carbonyl, was not formed. Sodium borohydride reduction of (20) furnished a mixture of the C7 epimers of (l6 )» "the composition being 1:1 (estimated by integration of the GLC peaks). Furthermore, the epimer with the lower retention time was shown by GLC to be identical to the axial -19-

ketol 0.6) obtained by complex hydride reduction of the enol ketone cn*

Cyclic ketones, particularly cyclohexanones, are known to 46 undergo ring enlargement with diazomethane or substituted diazomethanes*

However, rigorous examination (by IR, NMR, TLC and GLC) of the products obtained by treatment of (20) with nitrosoethylurethan in ethanolic 47,48 ether in the presence of sodium carbonate, or with ethereal 47 48 diazomethane, or with ethereal diazoethane, revealed only the presence of unchanged dione (20),

Hence although the studies of these bicyclo (4*3*l) decanes made no tangible contribution to the synthesis of the bicyclo (4*4*1) undecane (1), it was felt that the chemical and spectroscopic information derived from them would be invaluable when the desired analogous bicyclo

(4*4*1) undecane system came to be investigated.

As planned, (10) was reacted with methoxymethylenetriphenyl- 49, 50 phosphorane , initially at -70°, then at room temperature; the product was largely unchanged (1°), with small amounts of triphenylphosphine oxide (PhjPO) and two compounds, one slightly less polar, the other more polar, than (^°). Removal of unchanged (L0) by distillation,

and Of Ph^PO by chromatography, left material in which no enol ether

(25) could be detected. Several attempts under more forcing conditions

gave similar results, and it became evident that this route to (48),

(Scheme D), would have to be abandoned. -30-

Consequently, several alternative se

0-1) could not be prepared pure, (vide supra).

Ring-opening addition reactions between various nucle

(30), which have been hydrolysed by base to oc-ketoacetals (31). An analogous reaction between the ketoester (32) and ethyl diethoxyacetate

(29,R-Et) would produce the p-ketoester (33), convertible by base to (34) > "which possesses a four carbon side chain. Accordingly, esterification of the ketoacid (6) gave the ethyl ester (32), which hadV^j!^ 1699 (ketone) and 1737 cm.’"'*' (ester), and the following signals in the NMR spectrum: a quadruplet at 5*83tr

(2H, J= 7c/s,carbethoxy methylene protons), a triplet at 8*76U

(3H, J=7c/s, carbethoxy methyl protons, and a quartet at 7*54Ij

(2H, J=llc/s, Cy protons). Exposure of (32) and ethyl diethoxyacetate (29, R=Et) to the forcing conditions described for ( 29>R5SMe), ^6 afforded a product, the neutral fraction of which consisted of unchanged (32). The fractions soluble in bicarbonate and hydroxide solutions were both acidic, none of the ]3-ketoester (33) "being present.

Attention was then focussed on the second projected sequence to the ketoaldehyde (48)» (Scheme E). The enone (13) was prepared in very high yield by aldol cyclisation of the 1,5-dione (10) with methanolic potassium hydroxide. In larger scale preparations, however, the crude product from condensation of (4) with MVTC^? (vide supra)was heated at reflux overnight with base. Distillation separated unreacted (4) from the enone (13)» obtained in 65$ yield as a viscous pale yellow oil, which usually slowly solidified, and could be recrystallised only with some difficulty. Its 2,4-dinitrophenylhydrazone, C2oH26N4°4» was obtained as red needles. Thedp- unsaturated double bond in (13) gave rise to IR absorption at 1671 (C=0) and

1613 cm."1 (C«C), and to UV absorption a t ^ £ H 245 (£15000). -22-

A singlet in the MIR spectrum at 4*26 “X (ill) was ascribed to the vinylic proton.

Reduction of (13) with sodium borohydride afforded the allylic alcohol (35), distillation of which generated a small amount of a much less polar compound, probably the diene (36) (vide infra)•

Separation from (36), followed by cold recrystallisation, furnished

(35) as colourless needles with melting pt* 83-85°. Principal IR peaks were at 36IO (free hydroxyl), 1670 (C=C), 1024 (C-0) and 782 and 781 cm.”"1 (=C-H def.). The 60 Mc/s NMR spectrum exhibited a singlet at 8*25 X (hydroxyl proton), removeable by exchange, and a broad multiplet at 5*7 X (lH, carbinyl proton). In their 57 synthesis of thujopsene, Dauben and Ashcraft obtained the two allylic alcohols (37a) and (3Tb) by reduction of the enone (38).

The O'j olefinic proton, by coupling with the adjacent Cq carbinyl proton, gave rise to doublets at 4*63 X (J=l*3 c/s) and 4*48 X

(j=4»5 c/s) for the B-alcohol (3%) and the<£ -alcohol (3 7 b y e spectively. 58 In agreement with this, Finnegan and Bachman found that the pertinent signal in the allylic alcohol (39) at 4*55 X appeared to be only very weakly split (J * 1-2 c/s). On this basis, (39) was confidently assigned the quasi—equatorial configuration. In the case of (35), the 60 Mc/s spectrum contained a very slightly split (J ca. 1*5 c/s) doublet for the Cq olefinic proton at 4*67 X (Hi), suggesting that the hydroxyl group was quasi-equatorial, i. e*> ?' This was confirmed by the 100 Mc/s spectrum, in which the broad multiplet at 5*83© due to the

Ccj carbinyl proton became a triplet (j=8c/s) on irradiation of th.0

^8 proton signal. The value of J was of the order expected for 59 diaxial coupling be Ween the proton and the Cpo protons.

The Olefinic proton signal in this spectrum was a broad singlet

(half-band width 5c/s) at 4*76U, which was distinctly narrowed

(half~band width 3c/s) by irradiation of the C9 proton signal.

The mass spectrum of (35) showed no molecular ion, but a peak at

190 m/e (P-18) indicated facile dehydration.

That the contaminant in the distilled sample of the alcohol

(35) ( see p«2 2 ) was indeed the diene (3*5) was proved by the dehydration of (35) to (36) in high yield by polyphosphoric acid. The diene (56) exhibited identical TLC behaviour to the distillation contaminant.

Furthermore,UV absorption atAwax* ^73 mp.( £10300) confirmed the presence of a homoannular diene. In the IR spectrum, there was weak absorption at 3025(*OH), 1645 and 1590 cm.-1 (C=C), and strong absorption at 700 cm."1 (**C-H def.).

Since under basic conditions, only the enone (13) was formed by cyclisation of (10), it was decided to attempt the preparation with e,cidic reagents of the alternative product, the bicyclo (4»3»l) decenone (40). A mixture of unchanged dione (10) and enone (13), together with a trace of a less polar compound, was obtained on treatment 6 0 of (L0) withP -toluenesulphonic acid (PTSA) in refluxing benzene, or -24-

toluene, (in this case no starting material remained),or with concentrated sulphuric acid. The least polar compound was isolated in the first two cases, but IR spectroscopy demonstrated that neither was the desired enone (40). Only the enone (13) was found 62 to be present after prolonged reflux with boron trifluoride etherate.

These attempts to cyclise (10) served however to provide valuable expurience in the later cyclisation experiments on the ketoaldehyde

(48).

The next step in the route to (48), (Scheme E), was the preparation of the diene (41), This was obtained as a colourless, mobile liquid in 69$ yield after chromatography, by reaction of 49,63 (13) with methylenetriphenylphosphorane. Initially, the crude reaction products were chromatographed on silica, which achieved good

separation. However, one sample was completely isomerised during

chromatography to material with only very weak double bond absorption in the IR, and later runs were purified on neutral alumina. The

diene (42), synthesised in a similar manner by Sondheimer and fiA Mechoulom, was isomerised by concentrated hydrochloric acid to

the diene (43). Similar treatment of the pure diene (41) furnished material with IR absorption identical to that of the isomerised

chromatography product. The diene (41) appeared to undergo very

gradual atmospheric oxidation to the enone (13), and was not obtained

analytically pure. Its mass spectrum however indicated the correct molecular weight of 204, whileTXV absorption 243 (c

17800) and IH absorption at 3075, 1632, 1598, 894 and 865 cm."1

proved the presence of the diene grouping. Olefinic proton signals

in the NMR spectrum were observed as a sharp singlet at 4*17&

(lH, Cg proton), and a broad singlet at 5*30 “L (2H, exomethylene

protons)•

Hydrogenation conditions were next required to accomplish

selective reduction of the exocyclic double bond of (41). In their 65 synthesis of patchouli alcohol, Buchi and MacLeod employed a

similar sequence, viz., enone (45a) to olefin (45c) via the diene

(45b) # jhe hydrogenation step was carried out with Raney nickel 66 W2 in ethyl acetate. Be Grawand Bonner converted (46a) to (46e)

via (46b), the last stage being achieved with 10$ Palladium on

charcoal in ethyl acetate. After testing a variety of catalyst-

solvent combinations, smooth hydrogenation to the olefin (44) was

found to occur in 93$ yield with 5$ palladium on charcoal in ethyl

acetate. The resultant colourless, volatile oil was shown to have the

requisite double bond by its M R spectrum, which showed a doublet

at 4*77 U(JH, J«3»5 c/s) for the Cg olefinic proton, and no signal

due to a vinylic methyl group. This latter fact ruled out the

A8 isomer. The C9 methyl signal appeared to be at 8*93£> but its

splitting pattern was not clearly visible due to neighbouring peaks.

The XR spectrum was transparent in the 1650 cm. region (0=0 str.), -26-

but weak absorption at 875, 855 and 733 cm."1 was assigned to the trisubstituted double bond (=C-H deformations). The molecular weight of 206 was confirmed by mass spectrometry.

Ozonolysis of the olefin (44) yielded a complex mixture, spectroscopic examination of which did not detect the presence of the desired ketoaldehyde (48). Recourse was then taken to an alternative cleavage procedure. Treatment of (44) with osmium tetroxide in pyridine, followed by decomposition of the osmate ester with an aqueous pyridine solution of sodium bisulphite,^ furnished 94$ of the cis-7*8-diol (47), The purified diol, a colour less crystalline solid, Ci5H28®2» had two sharp bands in the IR spectrum at 3560 and 3&37 which were attributed to the Crj and

C8 free hydroxyl groups respectively. Important features of the llffi ft spectrum were a doublet at 6*59 £ (lH, J=9c/s, Cq carbinyl proton), a singlet at 8*20 t# (2H, hydroxyl protons), removed by addition of D2O, and singlets at 8«83, and 9*01, 9*08 t(l2H, Cq, C9 and C5 methyl protons respectively). The doublet expected for the C9 methyl group probably coincided partially with the C5 methyl signal.

The diol (47) was resistant to cleavage by excess sodium metaperiodate, in dry or aqueous methanol, even with long reaction 69 times. When lead tetraacetate was employed, the diol was smoothly cleaved to give (after removal of a trace of acidic material) 93$ of a neutral, pale yellow oil, which comprised mainly one compound,

* This value of J indicated that both the Cq Iljfdroxyl group and the C9 methyl group are equatorial. -27- the ketoaldehyde (48) • It exhibited. IR carbon/1 ab^rptitx ad ?.

(ketone) and 1727 cm. ^ (aldehyde). In addition, the C-H sir™ of

the formyl group was responsible for a weak band at 2700 cm. ^ A

doublet in the NMR spectrum at 0*37^ (iH, J=2c/s) confirmed the

presence of a formyl group with one adfacent proton, and a quartet

at 7.59T (j=llc/s), caused by the two protons,was also present.

The small acidic fraction was not investigated, but probably cc”3iso-

ed of the ketoacid (49)*

Cyclisation conditions were then required to convert the

ketoaldehyde (48) to the known bicyclo(4*4*l)undeceiione (50). it was

considered that acidic reagents would accomplish this more readily

than basic reagents, since the latter would tend to promote undesired

condensations involving the aldehyde grcup„ Indeed, treatment of tho

ketoaldehyde (48) with base furnished a complex of neutral compounds..

Similar aldol cyclisations to form bridged bicyclic compounds have 60 been executed by use of PTSA in refluxing benzene , or concentrated IT 6l 71 sulphuric acid 9 , These reagents afforded the dehydrated aldol 17 product. Others, such as acetic/dilute hydrochloric acids ,,''u gave

only the first-formed alcohol.

After several hours reflux with PTSA in benzene, (48) was

only partially converted into three much less polar compounds. The

incomplete reaction, and the complex product formed under these cond

itions, turned attention to the use of concentrated sulphuric acid.

Under these more drastic conditions, similar complex mixtures were -28-

however obtained, although no starting material was present.

The failure of this cyclisation step to afford smoothly the desired bicyclic skeleton could be attributed to three main factors. Steric hindrance by the gem-dimethyl group would discourage the desired reaction at C^. The alternative mode of cyclisation to form a bicyclo(5.0)decane compound (SchemeG) might also occur. Lastly, several other side reactions would be possible owing to the reactive nature of the formyl group. Table 1 Absorption frequencies in CCl^ of some trimethylcycloheptanone compounds

C-H def. 0=0 compound str. CH? C-Me CMe2 i C-Me 1 ] tetrahydroeucarvone 1704 1458 l^BO 1373,1367 (4) j R absorption due to R i 1 I CN 1703 1466 1380 1578,1367 ??55 (CaN) str. |

COpH 1701 1466 1380 1375,1365 3300-?500(bonded OH), 1710(0=0 )

COpEt 1699 1464 1389 1375,1366 1737(0=0), 1186(0-0 str.) j

, . . ~ u COpMe 1698. 1464 1388 1376,1366 1742(0=0 ), 1198(0-0 str.) ;

COCH5 1695 1464 1388 1374,1365 1721(0=0), 1352(0-11 def.)

CMepOH 1696 1464 1388 1374,1365 3ol0( free OH), 3530,33^ ' I (bonded OH) I Me CHCHO I698 n o t measure d 1727(0=0), 2710(C-E str.' of CHO)

* weak, broad absorption, decreasing on dilution (intermolecular hydrogen bond) Table 2 Absorption frequencies in CC1, of some bicyclo(4.3.1) decane compounds

C-H 1 def. 1 compound R C=0 str . CH other absorptions C-Me C-Me CMe? OH 1698 1454 1389 1376,1368 36l5(free OH),3445° (bonded OH) OTs 1702 1452 1390 c 1598,1494(ar.C=C), 1372,1340 and 1167, 1174(3=0 str.),906, 894(=C-H def.) =0 17 22 1456 1391 1378,1371 I696

1708 1453 1390 1376,1365 3030(=C-H),764,694a( def.)

H 1466 1385 1363 3625(free OH),3200- 3550 (bonded OH)

CH 1468 1382 1367 3ol5(free OH),3470b (bonded OH) a. in CS? b . centre of broad, weak absorption, intensity decreasing on dilution (intermolecular hydrogen bond) c . masked by SO^ str. absorption — "X* Table 5 Solvent dependence of the carbonyl stretching absorptions of the <: ;.ne (20) > 0 0 0

Solvent 0 11

11 v(V1-V 2)

Chloroform 1716 1691 25

Acetonitrile 1717 1692 25

Carbon tetrachloride 17?? 1696 26

Cyclohexane 1724 1699 25

* concentrations: ca. 0-015M in 0*5mm. cells

Table 4 Absorption frequencies in CCl^ of some bicyclo(5.4.0) undecane compounds

C-H def. compound R other absorptions C-Me and CMep I (13) =0 1462,1386,1377,1364 3020(=c-r),1671(0=0 ),1613(0=0 ) 1

OH 1458,1380,1359 36lO(free OH),l670(C=C),782

V. and 76la(C-H def.),1024(C-0) ! f O - i55> =CH2 1461,1384,1362 3075(=C-H),1632,1598(0=0),

V f <41> • 894 and 865(C-H def.), ^ (44) CH^ 1457,1376,1359 875,855,733(C-H def.)0 | ... 1 1 enol lactone(7) 1460,1389,1382,1358 1758(C=0),1662(C=C),1260 | (=C-0),also 1213,1174,1099,1065 diene (36) 14.62,1385,1365 3025(=C-H),l645w,1590w(C=C), 700s(C-H def.) ‘ diol (47) 1463,1382,1368 3637(free 2y0H),3580(free 370H)

a. in CS? b. thin film Table 5 Mass spectra of some compounds derived from •t e trahydroeucarvone

actual and loss of loss of other compoun obs. mol. wt. sidechain ch5 (15) peaks ketonitrile (5) 207 154(loss of 192 174(P-33) 53) ketoacid (6) 226 154(loss of 208(loss of K o0) 72) 209(loss of OH) 193(P-33) ketoester (32) 254 154(loss of 239 22l(P-33) 1 100) 209(loss of OEt) ; 236(P-18) ketoester (12) 240 154(loss of 225 209(loss of OMe) 86) 207(P-33) 222(P-18) diketone (10) 224 154(loss of 209 l9l(P-33) 70) ketol (8) 240a 154(loss of 207 189(P-33) 86) 222(loss of HpO) enone (13) 206 191 178(loss of CH?=CH2 or CO) allylic alcohol (55) 208a 175 190(loss of HpO) 91(c?h7+ ) diene (41) 204 189 91(c ?h 7+) olefin (44) 206 191 91(c7h ?+ ) enol lactone (7) 208 193 175(P-33) diene (36) 190 175 91(C?H7) diol (47) 240 225 222(loss of H ?0) 207(P-33) bicyclic ketol (16) 210 195 192(loss of HpO) ketotosylate (18) 364b 177 192(loss of PTSA) 172(PTSA) bicyclic diketone (20) 208 193 bicyclic enone (19) 192 177 j bicyclic diol (17) 212a i 179 194(loss of H 90)

a. no molecular ion observed. Losses were calculated xrom tne P—Id pear^. b. do., except P-172 peak. Table 6 jfeUi spectral assignments of tetrahydroeucarvone and some 2-substituted-2 ,6 ,6-trimethylcycloheptanones in CDC1-, 2

compound t y p e o f p r 0 t 0 n G ^ -Me C-Me ring and o C7 chain CH^ 7*55q. 9-oos 8-90d 8-45a J=llc/s 9* 05 s J=7c/s

R signals due to R CN 7*55q 9-05s 8*94s 8*49 J=ll»5c/s 9 *iOs COgH 7-57q 9 *05s 8 *98s 8.48a J=llc/s 9-13s j CO^Et 7-54q 9-02s 8*98s 8 *48a 5.48q,8*76t J=llc/s 9*12s both J=7c/s / 2\ 7-57q 9 -05s 8 «98s 8.48a 6.55s J=llc/s 9.13s Q i c v “ COCH5 7*55q 9-05s 9.01s 8 »48a 7.89s J=llc/s 9.16s CMe^OH 7-58q 9 *02s 8.97s 8-57a 8*52s(hydroxyl H) J=llc/s 9-12s 8.80s(methyl H) CE.MeCHO 7‘59q 9.05s 8.97s 8.49a 0«57h(J=2c/s, J=llc/s 9*12s formvl H) 8.86^(methyl H)

a. broad singlet centred on broader absorption b. possibly a doublet with J=7*5c/s, since a peak at 8*98, overlapped by the C^-methyl resonance, was detectable "able 7 I'M spectral assignments of some bicyclo(4.3.1)decane compounds in CDC1-, 3

compound type of proton

ring R methyl others methylene a OH 8-90s 8*48 5*50m (C7 carbinyl H) 8 »92s 7*80d(J=3.5c/s, C6H) 9 *00s 8*lsC(hydroxyl H) a j OTs 7 *58s(aromatic) 8*55 2» 22d , 2*66d(both J«8c/s) e-sastcp 4*’80m (C^ carbinyl H) 9-l3(C5) =0 8-80s(C.) 8*40£ 7.00se(C6H) 8-95s(C ) 7*2-7*5 complex (OgH) a enone (19) 8*90s 8-52 4 *13m(olefinic H) 9* 00s 7-45f (C6H) 7-72f (J oa. 2o/s,C? H)

9-03s 8 *66£ 6.32t(J=5c/s,C7 H) 9 -09s 6.458( ^ 2-50/3 ,0^ 11) 9 - 16s 8.J3s(both hydroxyl H)c

8*75s,6.77s 8.56* 8-3s(hydroxyl H)° (Cy and 01Q) 9*00s 9 *04s 9-13s a. singlet centred on broad absorption b. half-band width value proved the axial configuration of the group R (see discussion) c. signal disappeared on addition of D^O d. two protons ortho to the oxygen atom in the Ts group; the other doublet due to the remaining two aromatic protons e. very slightly split, possibly by long range coupling with or C s Me protons f. poorly resolved doublet Table 6 M iR spectral assignments of some bicyclo(5.4.0)undecane compounds in CDC1_

compound type of p r0 t 0 n

R methyl ring olefinic others methylene =0 8*84s(ci) 8*57a 4* 26s 8*98s ,9 *17s (c_)

8.67 a 4*67d , 5*7m(C carb- °h J ca.l*5c/ 9-07s,9-17s i „ i »> (c5) 8•25sc(hydrox yl H) 4-17s(C8 ) «CH2 9*00s(C ) 8 «66a 5 - 3 0 broads 9*C>5s,9*17s (exomethylc3ne H) (C5)

CH, 9*oos(c1) 8»72a 4*77d 3 J=3*5c/s 9-07s,9-15s (O5) 8-93d(C9)

enol lactone (7) 8-77s(C1) 8 • 28a 4-73s 8-9?s (C5 )

diol (47) 8-93s(c1 8-53a 6«59d(J=9c/s) and C d) 8•20sC(hydrox y 9*01s,9*08s yl H) (C5 } a. broad singlet, centred on broader absorption b. value of J showed the hydroxyl group-to be quasi-equatorial (see discussion) c. signal disappeared on adding D^O d. the other signal of the expected doublet was probably obscured by the other methyl resonances SCHEME A

SCHEME B

SCHEME C

COoH SCHEMED

OMe 10 25

SCHEM E E

0 0

11, R = COCl 13 14 12, R= CO?Me OTs

18 20

OH 0 21 22

\ z C02h

24 25 (RO),CHCOoR'2^* *^^21

2? R=CO?E t 28 29 CN

R' < > 0 (r o )2c h c o c h c o 2r (r o )2c h c o c h 2r

30 31 32 R = C 0 2Et

SCHEME F

A A H-Z + * ZCHCH CH( B B /CCLEt. Z= e.a. © / 2 A = B = C 0 2E t XC02E t

R C0CH(0Et)2 33 R=C02E-t 34 R= H 35 3G 37a,R=pOH 38 39 37b,R=ocOH

42 R = C gH ly 43 44

R 45 46 47 a , R - =0 , b > R= -CH2 , c , R=CH3 CHO 48 SCHEME G

48 -29-

ffXPERIMEmiL

Melting joints were determined On a Kofler microsooj e hot stage and are uncorrected, Boiling points are unconnected.

Routine infra-red spectra of liquid films and nujol mulls were recorded on Unlearn S.P. 200 and S.P. 200G instruments,

and solution spectra on a Unicam S.P. 100 double-beam spectro photometer equipped with an S.P. 130 sodium chloride prism-grafcing

doable monochromator operated under vacuum conditions.

Ultra-violet absorption spectra were determined on a

Unicam S.P. 800 s.pectrophotometer. Nuclear magnetic resonance

spectra were recorded on a Perkin-Elmer R.S.10 (6QMc/s) spectrometer, using solutions in deuterochloroform with tetramethylsilane as

internal reference.

Gas-liquid chromatography was carried out on Pye-.trgon and

Perkin-Elmer Fll Gas Chromatographs. Mass spectra were recorded on an A.E.I. M.S.9 spectrometer.

Routine and preparative thin layer chromatography (TLC)

employed Kieselgel G (Merck) silica, with 20$ or JOfo ethyl acetate

in light petroleum as the solvent system. Chromatoplates were

developed with iodine vapour.

All organic extracts were dried over anhydrous magnesium

sulphate, and unless otherwise stated, previously washed with brine. -30-

Bucarvone JL2)

Eucarvone was prepared from c&rvon© (via carvon© hydro- 26 27 bromide ) by the method of Wallaeh . Yield 62^, b.p, 44°/0*05nm., 70 ^ 1*5050* (Lit. , b.p. 82—84°/8nm., 11^^1*5056)•

Te trahydroeuoarvone (4)

Hydrogenation of euoaimrone by the method of Barnes and 2 8 Houlihan yielded totrahydroeuoarvone. Yield 84#, b.p. 39°/0,4mm.,

up 1*4565, semicarbazone m.p, 171-173° (3x aqueous EtCH).

(lit. 7 4 b.p. 42-43*5°/0.4 mm., np^l*4548, semicarbazone m.p,

185-187°).

• IR, table 1 ; NMR, table6 «

2-(jB-cyanoethyl)-2,6*6-trimethyIcycloheptanone (5)

A solution of tetrahydroeucarvone (30*8 gm.; 0*2M) and

Triton B (40^ aqueous solution? 4 ml.) in ethanol (150 ml.) was warmed 29 to 50 » and acrylonitrile (21 *2gm.; 0*4M) was added dropwise with

stirring over l£ hr. After being stirred for 5 hr. at the same

temperature, the dark yellow mixture was acidified (dilute HCl) and

extracted with ether* The organic extracts were washed (brine),

dried, and evaporated, and the residue distilled, yielding unreacted

tetrahydroeucarvone (8*4gn*)> ond (5) as a colourless oil which slowly

solidified (l7*2gm.; 57$ based on consumed tetrahydroeucarvone), b.p, 104°/0*3mm., m,p# 44.5-46.0° (ligit petroleum 40-60°),

Found; 0,75*13 ; H, 10*18 ; N,6*75* C13& 23ON requires 0,75*32; g,

10*21; N, 6*76. Semicarbazone, m.p. 46*0**46*5° (aqueous EtOH)•

IR, table 1 ; MR, table 6 ; mass spectrum, table 5 •

2rii^parbo}Qrethyl)-2.6,6~trimethylcycloheotanane (6)

The ketonitrile (5), (16*56 gm.; 0*08M), dissolved in dioxan (200 ml.), was added to concentrated hydrochloric acid (80 ml.) in water (80 ml.), and the mixture stirred under reflux for 3 days.

The ether extract was washed with saturated sodium bicarbonate solution*

Ether extraction of the acidified bicarbonate layer yielded (6) as a brown oil which usually slowly solidified (17*7 gm.; 98$), The very pale yellow oil, b.p. ca. 125°/0*4 mm., obtained by distillation, solidified, and was recrystallised from pentane as a colourless solid, m.p. 57-58°.

Found; C, 69*40; H,9*31* O13H22O3 requires 0,68*99? H, 9*80.

IR, table 1 ; MR, table 6; mass spectrum, table 5 *

1.5.5-trimethyl-8-oxabicyclo(5.4.0) undec-6~en-9-one (7)

Sodium acetate (500 mg.) was added to a solution of the ketoacid (6), (ll*3 gm.; 0*05M), in acetic anhydride (500 ml.) and the 3 0 P R mixture refluxed for 4 hr. 9 D Excess acetic anhydride was removed by distillation under reduced pressure and the residue dissolved in ether. Distillation of the filtered, evaporated ether solution afforded

(7) as a colourless oil, b.p. ca. 130°/0*8 mm., (8*0 gm.; 1 7 $ ) . A sample was redistilled (b.p. 70-80°/°*05mm.), and on standing became a colourless crystalline solid, too soluble to be recrystallised even from the less polar solvents. It also appeared to be hygroscopic, and satisfactory analysis figures could not be obtained*

IR, table 4 » MR, table 8 ; mass spectrum, table 5 •

Attempted preparations of the 1.5-dione (10) a) Treatment of the enol lactone (7) with methyl magnesium iodide . \ 31 i) 0*5 M excess of Grignard reagent added to enol lactone (7).

A suspension in dry ether (5 ml.) of MeMgl (3*75mM),

prepared from magnesium (91 mg*) and methyl iodide (533 mg.) was added

dropwise with stirring to a solution of the enol lactone (7)(520 mg. ;2*5uiM)

in dry ether (25 ml.). After 1 hr. stirring, the mixture was acidified

(10^ dilute HCl), and extracted with ether. The ethereal layer was

washed with saturated sodium bicarbonate solution to separate neutral

material (184 mg.) from ketoacid (335 mg,). The former was shown by

IR and TLC to be unreacted (7)» 3 2 ii) enol lactone (7) added to a 5 M excess of Grignard reagent *

An ethereal solution of (7)» (520mg.; 2*5 nM), was added over

1 hr. dropwise with stirring tojMeJfgl (l5mM). After being stirred

at reflux for 3 hr., extraction as in (i) yielded acidic material

(52 mg.), and a neutral pale yellow oil (445 mg.), which exhibited strong

carbonyl (1700 cm.”^) and hydroxyl (35°0 cm. ) absorption in the IR, and

contained three compounds, in amounts decreasing with polarities (TLC).

Repetition of the reaction without heating to reflux, yielded an identical

product (435 mg.). Chromatography of both products on silica (30 gm.),

eluting with light petroleum/EtOAc mixtures afforded: the least polar -33-

component (21 mg.; eluted with petrol), which showed weak IE

carbonyl absorption; the middle component, (121 mg.; eluted with

petrol/ 1 and 2$ EtOAc), identical (lR,TLC) to 1,5-dione (lo) from

the reaction between HVK and (4); and the most polar component,

a yellow oil, (587 mg.; eluted with petrol/ 5-10$ EtOAc).

Distillation of the last gave a colourless oil, b.p. 90-95°/0*5mm.,

shown to be 2-(y~hydroxyisoamyl)-2,6,6-trimethylcycloheptanone (8)

from its IR, NMR and mass spectra (tables 1 ,6 , and5 respectively). b) Treatment of the ketonitrile (5) with methyl Grignard reagent

A mixture of the ketonitrile (5), (2*07 gm.; 10 mM), and 33 MeMgBr (20 mM) was refluxed for 2 hr. Extraction as in (a) yielded

acidic material (30 mg.) and a neutral yellow oil (1*7 gm.), which

contained considerable unchanged (5), (IR, TLC), together with a

compound of similar polarity and another much less polar (TLC)•

This material was not further examined due to the small extent of

reaction.

Ketonitrile (5), (1*055 gm.; 5 mM), on treatment with 33 MeMgl (40 mM) formed a neutral product (952 mg.) showing a doublet -1 in the IR carbonyl region at 1720 and 1700 cm. • TLC analysis

indicated only a trace of starting material, and four or five other less

polar components. This complex mixture was not further investigated.

c) Treatment of the ketonitrile (3) with methyl lithium Following tile procedure used in th© f* oil owing e xp© aritnen t r

methyllithium (60 mM) and the ketonitrile (5 ), (l. 552 gm.; 7*5 mM), 33 were stirred under reflux for 3 hr. The neutral product (1275 mg.)

exhibited strong carbonyl (1710 cm.~^) and weak hydroxyl (3500 cm. ^)

absorption in the IR, and consisted of three main compounds (TLC).

Separation by preparative TLC (thickness 1 mm., eluant light petroleum/

30$ EtOAc) yielded from the predominant central band, reasonably pure

1,5-dione (10), (492 mg.; 29$), identical (IR, NMR) with that obtained

by condensation of MVK with (4 ), (vide infra). Hater runs however gave

lower yields of material containing much less (10), and more of a less

polar compound. For example, the product (1*34 gm.) from the

ketonitrile (2*58 gm*; 12*5 mM) afforded after chromatography only

0*18 gm. (6 *5$) of (10). d) Treatment of the ketoacid (6 ) with methyllithium

Methyl iodide (2*84 gm.; 20 mM) was added dropwise over

30 min.to a stirred suspension of lithium (280 mg.; 40 gm. A) in dry

ether (10 ml.). The mixture was stirred under reflux for 1 hr., by

which time almost all the lithium had dissolved. The colourless cloudy

solution was filtered through glass wool into a pressure-equilibrated

dropping funnel and added over 5 min. to a stirred solution of the 34 ketoacid (6 ), (1*13 gm.; 5 mM), in dry ether (15 ml.). (N‘2 atmosphere

throughout) * After being heated at reflux for 2 hr., water was added

carefully to the cooled solution. The ether extract was washed with

saturated sodium bicarbonate solution to separate unchanged acid (IO46 mg. from neutral material (50 nag*), shown by TLC to be a mixture of three

compounds. Further attempts at this reaction gave similar results.

Owing to the low yield, and complexity of tide product, the material

was not further examined. e) Attempted preparation of pure ketoacidchloride ftl)

a) thionyl chloride.

The ketoacid (6), (11*3 gm.; 0*05 M), and thionyl chloride

(12-1 gm.; ca. 20 ml.; 0*1 M) were heated to reflux for 3 hr. Excess

thionyl chloride was distilled under reduced pressure (water pump).

Distillation afforded a pale red viscous oil vhich partially solidified

(7*7 gm.), b.p. 120-122°/0*15 mm* The distillate contained the title

compound (odour% IR 1700, 1800 cm.“^), but the predominant component

was the enol lactone (7), (IR 17&0, 1660 cm. "*")• TLC also indicated

the presence of two compounds, the less polar being (7 ).

b) Oxalyl chloride (5ml.; 0*06 M) was added to a solution of the

ketoacid (6 ), (2*26 gm.; 0*01 M), in dry benzene (20 ml.). The

mixture was stirred for 8 hr., and left overnight. Benzene and excess

o&alyl chioxide were distilled under reduced pressure (water pump)»

The residue was shown by IR and TLC to be a mixture similar to that

obtained in (a).

2- (ft-carboaa thoxye thyl)-2 ♦ 6 ♦ 6-trime thyl cycloheptanone (12 )

Ihe ketoacid (6), (2*26 gm.; 0*01 M), dissolved in dry Et20 -36-

(20 ml.), was treated dropwise with ethereal diazomethane until a

yellow colour persisted. The solution was left overnight, filtered

to remove polymer, and evaporated, yielding (12) as a pale yellow

oil (2*24 gm*J 93$)* Distillation afforded a colourless oil,

b.p. 106-107°/3mm.

Found: 0,69*21; h, 9*40. C^B^O^ requires C, 69*96; h»10*07*

IR, table 1 ; RMR, table 6 ; mass spectrum, table 5 • f) Treatment of the ke toes ter (12) with dimsyls odium

A 50$ mineral oil dispersion of sodium hydride (1*2 gm.; ca,

25 mM) was washed thoroughly with light petroleum 40-60°, and the washings

decanted. The flask was then filled with nitrogen, and dimethyl-

sulphoxide (33MSO), (25 ml.; distilled from calcium hydride) was added

dropwise with stirring. After heating the mixture at 70-75° for 1 hr.,

and cooling, there was obtained a pale grey solution of the sodium salt 36 z x of IMSO. To 15 ml. of this solution (ca,15 mM) was added dry THF

(15 ml.),and, with cooling in an ice bath, the ketoester(l2), (2 gm.;

8*3 mM), dropwise over several minutes. After being stirred for 30 min.

at room temperature, the solution was poured into water (90 ml.), acidified

to pH 3-4 (dilute HCl), and extracted with chloroform. The organic

extract was washed (3 x water), dried, and evaporated, leaving a yellow

oil (2*36 gm.). The large amount of acid present (1*8 gm,) was removed

by washing with saturated sodium bicarbonate solution, leaving neutral

material (473 mg.). This consisted of two compounds, both less polar than the ketoester (12), Since the material showed no carbonyl absorption in the IR, it was not further examined,

2-(31 -ketobutyl)-2«6.6-trimethylcycloheptanone (10 )

Treatment of tetrahydroeucarvone (19*86 gm,; 129 mM) in dry ether (100 ml,) with potassium hydroxide (0*84 gm,; 15 mM) in dry ethanol (15 ml.) and methyl vinyl ketone (4*90 gm,; JO mM) by 37. a method similar to that for the preparation of (L3) (vide infra), yielded on distillation unchanged (4), (10*01 gm,), and a fraction

(7*40 gm.) with b.p, 120-124°/0 •8 mm. This material contained mainly

(10), with small amounts of (4) and (13), and was chromatographed on silica (370 gm.), Tetrahydroeucarvone (0*3 gm*) was eluted by light petroleum/5-50$ benzene, the enone (13), (0*82 gm.),by benzene, and

(I^Dj (5*48 gm.; 39*5$ based on consumed (4)), by ben2ene/l0~50$ chloroform. The dione (1 0) was obtained as a colourless oil by distillation, b.p, 117-120°/0*3 mm.

IR, table 1 ; RMR, table 6 ; mass spectrum, table 5 •

The mono-2,4-dinitrophenylhydrazone was crystallised from ethanol as a yellow crystalline solid, m.p, 113-114°*

Found: C, 59*58; H, 6*92 ; N, 13*95* 820^28^485 requires 0,59*39; H,6*

N, 13*85 Yggl4: 1699 cm."1 l,5t5-trimethylbicyclo (4.3.1) decan-7-ol-10-one Q6)

A suspension of lithium aluminium hydride (2*56 gm.; 67 mM) in dry THF (100 ml.) was added dropwise to a stirred solution of t butanol (14*94 gnu; 201 iriM.) - 'Thxa t^ufsporvaion ©f 1 i’fchinm al.umiw-iA.ini 38 tri-(t-butoxy) hydride was added dropwise over 1*5 hr. in a nitrogen atmosphere to a stirred solution of the enol lactone (7)* (10*0 gm,; 25 48 mM) , in THF (100 ml.) at -70° (acetone - dry ice bath). The mixture was allowed to warm to room temperature while being stirred overnight. After addition of brine, acidification (dilute HCl), and ether extraction, the ether layer was washed with saturated sodium bicarbonate solution, brine, and dried and evaporated, leaving a pale yellow oil (7*30 gm.). Ether extraction of the acidified bicarbonate layer afforded ketoacid (6), (2*39 gm.). The neutrals consisted of three compounds, the most polar being (16), and the next most polar, which varied in amounts in different runs, probably being unchanged (7)

(similar Rf). Chromatography on silica (360 gm*)> eluting with light petroleum/EtOAc mixtures, afforded (16) as a pale yellow oil (3*9 gm.;

39$), eluted with light petroleum/20-30$ EtOAc. Distillation furnished axial (16) as a colourless oil, b.p. 80-90°/0*l mm., which slowly solidified to a colourless crystalline solid, m.p. 71-76°. Later runs afforded axial-equatorial mixtures which became colourless liquid-solid mixtures on distillation, final m.p. ca. 70°.

IR, table2 ; MR, table 7 » mass spectrum, table 5* l«5.5-trimethylbicyclo (4*3*1) decan-7,10-diol (L7)

A solution of the ketol (16), (210 mg.; 1 mM), in dry THF

(7 ml.) was stirred 24 hr. under reflux with lithium aluminiumhydride -39-

(114 nig.; 2 M excess). Acidification (dilute HCl) and. ether extraction gave a viscous pale yellow oil, consisting of diol (if)* unchanged (16), and traces of a less polar compound. Separation by preparative TLC (thickness 0*8 mm., eluant: light petroleum/60$

EtOAc) afforded pure diol (62 mg.; 29$)» m.p. 91-97° (petrol 60-80u).

Found: C,73*45; H , 11*50. ^ 5^24^2 re^uires C,75*54# &# 11*39«

IR, table 2 ; HMR, table 7 » mass spectrum, table 5 •

* The products from reductions carried out under several different conditions always contained unchanged (16). After a

45 hr. reflux with excess lithium aluminium tri-(t-butoxy) hydride, the product contained only a trace of (16), and also several less polar compounds in small concentrations.

1«5«5-trimethyl-7-tosyloxy-bicyclo(4.5.1) decan-10-one (18)

To the ketol (16), (420 mg.; 2 mM), dissolved in dry pyridine ( 2ml.), was added a solution of tosyl chloride (4®Q mg.; slight excess of 2 mM) in dry pyridine (2 ml.). After being left at room temperature for 14 days (the product after 4 days contained an appreciable amount of unreacted (16)), brine was added, the mixture acidified (dilute HCl) and extracted with ether. The ether extract was washed (brine), dried, and evaporated, leaving (18) as a yellow oil which slowly solidified (437 mg,; 60$). A colourless crystalline solid, m.p, 11^-115°, was obtained by recrystallisation from ethanol.

Found: C, 66*12; H , 7*58- C20H28O4S requires C,65.91;B*7-74* -40-

IR, table 2 # NMR, table 7» mass spectrum, table 5 •

1 # 5»5-trimethylbicyclo (4.3.1) dec-7-en-10-one (19) a) from ketotosylate (18)

The ketotosylate (18), (91 mg.; 0*25 mM), was added in small portions over 5 min. to a stirred solution of sodium (7 mg.) in dry ethanol (2 ml*). The mixture was stirred under reflux for 2 hr., cooled, neutralised with acetic acid, diluted with water and extracted with light petroleum 60-80° (3X)f”^ Ether extraction of the aqueous layer afforded unchanged (18), (5 mg.). The petrol extract was washed with saturated sodium bicarbonate solution, brine, and dried. Evaporation of solvent under reduced pressure yielded

(19) as a volatile colourless oil with a camphoraceous odour (33 mg.;

73$ based on consumed $ ) UV A MaxP^Ogmp. (e3600).

IR, table 2 ; MR, table 7 ; mass spectrum, table5 • b) attempted preparation from ketol (16)

A solution of the ketol (16), (105 mg.; 0*5 mM), in dry benzene (2 ml.) was refluxed on the steam bath for 2 hr. with excess poly-phosphoric acid (treatment at room temperature caused incomplete reaction). Addition to ice, followed by ether extraction, afforded a volatile yellow oil (75 mg.), containing a rather non-polar compound

(large Rp in hexane), and several more polar components (TLC).

Routine IR showed carbonyl absorption at 1705 cm.“^. Chromatography on silica (eluantspentane) afforded the predominant component as a -41-

colourless volatile oil (18 mg,), exhibiting no carbonyl absorption in the IR*

1 >3>5-trimethylbicyclo (4.5.1) decan-7,10-dione (20)

A solution of ketol (16), (2*10 gnu; 0*01 M), in acetone / \ 42 (50 ml.) was treated for 30 min. with excess Jones reagent at 0 .

Methanol was added to destroy excess oxidant. After addition of brine, extraction with ethyl acetate furnished (2 0 ) as a colourless oil (l«89 gm.; 91$), b.p. 60-70°/0*02mm.

IR, table 2; MR, table 7; mass spectrum, table 5 •

GLC columns 1$ PEGA, Temp, s 125°• F.R. 48 ml./min. R.T. s 5 min.

Mono-2,4-dinitrophenylhydrazone, m.p. 188-191° (EtOH).

Found: C, 58*80 ; E, 7*03* Q19K24N4O5 requires 0,58*75» H, 6*23*

Attempted oyclisation of Icstoacidfioteo dione (20)

Tetralin ( 5ml.) and p-toluene sulphonic acid (100 mg.) were heated together to reflux for 1 hr. in a Dean and Stark water separator ^ The ketoacid (6 ), (113 mg.; 0-5 mM), was added, and the mixture was refluxed for 2 hr. Brine was added, and the ether extract was washed with saturated sodium bicarbonate solution and brine, and dried. The neutral product (ill mg.), a brown oil, had IR absorption at 1680 cm,'”’*’ (0=0), and exhibited marked streaking in TLC. After distillation however, TLC showed it to be a mixture of four compounds.

The most polar of these had an Rf value similar to that of (30) f hut the -42-

mixture was not further investigated.

Attempted cyclisation of the ketoest^r (12) to the dione (20)

m u 45 The procedure was based on a reaction by Marshall and Scanio.

A 51$ dispersion of sodium hydride in mineral oil ( 0*6 gm*) was washed with dry benzene and dry 1,2-dime thoxy ethane (EME) and added to a solution of the ketoester (12), (780 mg.; 3*25 mM), in

DME (10 ml.). After 14 hr, reflux, acetic acid (1*5 ml.) and dry ether (10 ml.) were added carefully to the cooled solution. The ether extract was washed with saturated sodium bicarbonate solution to separate neutral material (72 mg.) from ketoacid (6), (571 mg.).

The former contained some unchanged ester, but none of the desired dione.

Treatment of the dione (20) with methyl magnesium iodide (MeMgl)

A solution of the dione (20), (104 mg.; 0*5 mM), in dry ether (3 ml.) was added dropwise with stirring to an ethereal suspension of MeMgl (0.65 mM). After being refluxed for 1 hr., water was carefully added. Ether extraction yielded a pale yellow oil (99 mg.), containing mostly unchanged (20), and two more polar compounds, -*43~

probably the C7 epimeric ketols (9), (TLC). Repetition of this procedure with a larger excess of MeMgl (2*5 mM) gave a product

(95 mg.) containing in addition to the above, a highly polar compound, probably the diol (22). The products were combined and stirred under reflux for 16 hr# with MeMgl (2*5 mM), affording material (174 mg.) of largely unchanged composition (TLC). Chromatography on silioa

(9 gm.) using light petroleum/EtOAc mixtures, afforded unreacted dione

(20), (48 mg.; eluant petrol/l and 2$ EtOAc), a mixture of two compounds, probably the epimeric ketols (9) (66 mg.; eluant petrol

/5-10$ EtOAc), and the diol (22), (40 mg.; eluant petrol/20-30$

EtOAc), which after distillation (b.p. 85-95°/0*l mm.) and re crystallisation from light petroleum 60-80° became a colourless crystalline solid, m.p. 85-114°•

IR, table2 ; NMR, table 7.

Conversion of the dione (20) to the ketoacid (6)

A solution of the dione (20), (52 mg.; 0*25 mM), in methanol

( 5ml.) w«s heated to reflux on the steam bath for 30 min. with excess methanolic potassium hydroxide. Methanol was then allowed to evaporate from the straw-coloured solution, and the residue was acidified (dilute HCl). The ether extract was washed with brine, and saturated sodium bicarbonate solution. Ether extraction of the acidified bicarbonate layer gave an acidic yellow oil (54 mg.; 96$).

The distilled product was identical (lR» HMR, TLC) to the ketoacid (6)* -44-

Sodium borohydride reduction of the dione (20)

The dione (20), (52 mg*; 0»25 mM), dissolved in methanol

(3 ml*) was treated with sodium borohydride (10*5 mg*; 0*275 mM),

and the mixture was stirred overnight. Water was added, and ether

extraction yielded a pale yellow oil (50 mg*; 95/^) which consisted

of two components (TLC). GLC Column: 1%* PEGA Temp: 125°.

F.R. 40 ml./min. R.T. 16*0min. (identical R.T. to axial (l6 ) from reduction of enol lactone (7)) and 37*6 min. (hence equatorial ketol

(16 ) ) • The composition of the mixture (by integration of the peaks) was 1 :1.

Treatment of the dione (20) with diazoethane

a) N-nitroso-N-e thy lure than method^ * ^

Anhydrous sodium carbonate (3 mg.) was added to a solution of the dione (2*3), (104 mg.; 0*5 nil), in ethanol ( 1 ml.) and ether 73 (l ml.). N-nitro so-N-ethylurethan (oa235 mg.; 2 mM; prepared from

N-ethyl carbamate74 ) was added dropwise over 5 min., and the mixture

stirred for 30 hr. After acidification (dilute H2SO4), and 26 hr. stirring, brine was added, and the solution extracted with ether.

The ether extract w%s washed (brine), dried, and evaporated. Distillation of the residue gave a colourless oil (73 mg.), b.p. 65~75°/0'l mm., which consisted solely 9f unchanged (2Q, (IR, TLC, NMR and GLC). fc) N-nitroso-N-eth.vlurea method

Excess N-nitroso-N-ethylurea was added to a mixture of 40/ -45-

aqueous potassium hydroxide (7 ml.) and ether (7 ml.) The resultant yellow organic layer was separated and left for 3 hr. over solid

KOH with cooling. It was then added dropwise with stirring to a solution of the dione (20), (52 mg. j 0*25 mM), in dry ether (2 ml.).

After being stirred for 2^ hr., the ether solution was washed (brine), dried, and evaporated, affording a pale yellow oil (52 mg.).

IR, TLC and GLC showed it to contain only (20).

47 Treatment of the dione (20) with diazomethane

The dione (20), (31 rag.; 0*15 nM), dissolved in dry ether

(l ml.) was treated with excess ethereal diazome thane. After being stirred for 21 hr., the solution was evaporated, yielding a pale yellow oil (30 mg.) consisting only of starting dione (IR, TLC, GLC).

Treatment of the dione (10) with triphenylphosphinemethoxymethylene

The method used was that for the preparation of methoxy- 75 • me thylenecy clohexan© . 76 A suspension in dry ether of phenyl lithium (5 mM)- prepared from bromobenzene (0»78 gm.) and lithium (70 mg.) - was added dropwise with stirring in a nitrogen atmosphere to a suspension of triphenyl methoxymethylphosphonium chloride (l* 70 gm.; 5 mM) in dry ether (15 ml.). The resultant cloudy red solution, consisting of 5 ° a suspension of methoxymethylenetriphenylphosphorane, was stirred for 10 min., cooled to-70° (dry ice- acetone bath) and the dione (1 0 ), -46-

(896 mg.; 4 mM), in ether ( 2 ml.) was added dropwise. The mixture was stirred for 15 min. at -70°t snd 19 hr. at room temperature.

The cloudy off-white solution was filtered, the filtrate evaporated, and the residue dissolved in acetone, diluted with a large volume of light petroleum 40—60°, and carefully filtered. Repetition of this process removed most of the triphenylphosphine oxide, and yielded a yellow oil (ca. 1 gm.), which was freed from starting dione (10) by distillation. The residue (220 mg.) consisted of a trace of unchanged (10), Ph^PO* a compound slightly less polar than (10), and another highly polar (remaining on baseline in TLC). Removal of Ph^PO by chromatography left a brown oil, IR examination of which did not detect the presence of the desired enol ether.

Similar products were obtained by adding the dione at room temperature, or heating the mixture to reflux for 2 hr.

Attempted preparation of 1,1-dicarbethoxycyclopropane (27, R^COoEt) 53 The method of Doxandlfoder afforded on distillation in the range 200-220° a colourless oil. (Lit. title compound b.p.

214-2l6°/748 mm.; AOfo). Both its IR and MIR spectra- showed it to be unchanged diethyl malonate (b.p. 199°)* In particular, the M R spectrum showed no signal due to cyclopropyl protons.

Attempted preparation of l-carbethoxy-l-cyano-cyclopropane (27, R-CB) 54 5 4 The method of Perkin, modified by Jones and Scott, ^ -47-

yielded from ethylcyanoacetate (100 gm.; b^p, 206°) after steam

distillation a colourless oil (13*5 gm*)* Fractional distillation

(Lit. title compound, b.p. 212-216°; 64*4 gm*; 76fo) afforded one main fraction; b.p, 206-210°, shown by its IR and M R spectra to

be unreacted ethylcyanoacetate.

carbe thoxyethyl)-2 ,6 .6-trime thylcycloheptanone (52)

A mixture of ketoacid (6 ), (6*78 gm.; 0*03M)» ethanol

(100 ml.) and dilute sulphuric acid (60 ml.) were stirred for 1 hr.

After addition of brine, the ether extract was washed with saturated

sodium bicarbonate solution, (ether extraction of the acidified

bicarbonate layer affording unchanged (6 ), (1*6 gm.)), brine, dried

and evaporated, yielding (32) as a pale yellow oil (5*6 gm.; Q&fo based on consumed 4)* Distillation gave a colourless oil, b.p. 96°/0.08 mm.

IR, table 1 ; KMR, table <5 ; mass spectrum, table 5 •

Attempted condensation of ketoester (52) and ethyl diethoxyacetate

The forcing conditions described for condensations with 56 ethyl dimethoxyacetate were employed.

A mixture of the ketoester (32), (2*54 gm.; 10 mM), and ethyl diethoxyacetate (2.9,R=Et), (1*76 gm.; 10 mM) in dry benzene

(10 ml.) was added to a, suspension of sodium ethoxide (from 240 mg,

sodium and 7*5 ml* dry ethanol, followed by evaporation of solvent) -48-

in dry benzene (lO'ml.). The benzene—ethanol azeotrope was removed over 5 hr. from the refluxing, stirred solution via a fractionation column. Gold dilute acetic acid (5 ml.) in water (5 ml.) was added to the cooled mixture. The ether extract was washed with saturated sodium bicarbonate solution, saturated sodium hydroxide solution, and brine, and dried. Removal of solvent left a yellow oil (1*3 gm.) consisting of unchanged (52), (IR, TLC). Ether extraction of the acidified bicarbonate and hydroxide extracts gave in each case acidic material (IR, TLC; 0*6 gm. in all),

1.5* 5-trimethyitficyclo( 5 *4*0) undec-7-en-9-one (15) a) from tetrahydroeucarvone.

Potassium hydroxide (455 rag*; 8*12 mM) in dry ethanol (5 ml.) was added to tetrahydroeucarvone (16*69 gm.; 108*3 mM) in dry ether

(100 ml.). The mixture was cooled to 0°, and freshly-distilled 37 methyl vinyl ketone (3*79 gm.; 54*15 mM) in ether (25 ml.) was added over 1 hr. dropwise with stirring. The ice-bath was removed, and after being stirred for 1 hr., the pale red solution was poured on to ice, neutralised (concentrated HCl), and extracted with ether.

The product, a dark yellow oil (ca. 19*5 gm., comprising tetrahydro eucarvone, 1,5-dione, and (13)) was dissolved in methanol (100 ml.) and refluxed overnight with potassium hydroxide (5 gm*) in methanol

(50 ml.). Ether extraction (assisted by the addition of brine), removal of solvent, and distillation afforded tetrahydroeucarvone -49-

(9-87 gnu), b.p, ca. 35°/0*15 mm,, and 0-3), (3*96 gnu; 65^ based on consumed tetrahydroeucarvone) as a pale yellow oil which slowly solidified* b.p. 92-94°/0*15 mm. Recrystallisation and cold filtration frofci pentane furnished with difficulty a colourless crystalline solid, m.p. 67*5-68*5° TTVAlso? 245 (£ 1500 ^

The 2,4-dinitrophenylhydrazone was crystallised frdm ethylacetate as red needles, m.p, 210-213°.

Pounds C,62*27? H, 6*79? N, 14*58. C2oH26N404 squires C,62*16; H,

6*78; N, 14*50.

IR, table 4 » NMR, table 8 ? mass spectrum, table 5 * b) from 2-(5* -ketobutyl)-2«6 ,6-trimethylcycloheptanone (10)

Potassium hydroxide (260 mg.) in methanol (3 ml.) was added to a solution of the 1,5-dione (10), (224 Kg.; IraM), in methanol

(2 ml.), and the mixture refluxed for 4 hr. on the steam bath. Brine was added, and the ether extract was washed (brine), dried and evaporated, yielding (1 3 ) as a pale yellow oil (198 mg.; 96fo) identical to that from (a), (IR,TLC).

1 * 5 * 5-trimeth.ylbicyclo( 5 >4* 0)undec-7~en-9 E-qi P 5 )

A solution of the enone (13)> (206 mg,5 1 mil), in ethanol

(15 ml.) was treated with sodium borohydride (38 mg.; 1 mil), and refluxed 13 hr. with stirring. Water was added to the cooled solution, and the ether extract was washed (brine), dried and evaporated to yield

(35) as a pale yellow crystalline solid (208 mg.; 100)0* This -50-

material was soluble in pentane at room temperature, but colour less (35) obtained by distillation was recrystallised and filtered cold from that solvent, affording colourless needles, m.p, 83-85°.

Some diene (formed by dehydration) was present in the distilled sample (TLC), and was removed by preparative TLC (eluanti light petroleum 40-60°).

Prom its HMR spectrum (table8) the hydroxyl group was seen to be

/ n57,58 quasi-equatorial (B).

IR, table 4; mass spectrum, table 5 •

1,5,5-trimethylbicycl*(5.4.0)undeca-7,9-diene (36)

The uns.aturated alcohol (3!$» (208 mg,; lnM), dissolved in dry benzene (3 ml.) was stirred for 1-J- hr. with excess polyphosphoric acid. Addition to ice, followed by ether extraction, yielded v EtCE (3 6) as a colourless oil (178 mg.; 94/^)* A Max 273 mp (f 10300)

IR, table 4 J mass spectrum, table 5 •

Attempted c.vclisation of the dione (10) to 1,5,5.7-tetramethylbicyclo

6 0 a) refluxing with p-toluenesulphonic acid (PTSA) in benzene

A solution of PTSA (500 mg.) in dry benzene (15 ml.) was heated at reflux for 2 hr. in a Dean and Stark water separator. The dione (10, (448 mg.; 2mM), dissolved in dry benzene (3 ml.) was added -51-

to the cooled solution, and the mixture refluxed a further 2g- hr*

After neutralisation with solid potassium carbonate, the solution was left overnight, and filtered. The filtered solid was washed with benzene, and the combined benzene extracts were washed, dried, and evaporated, yielding a brown oil (434 nig.). The product contained mainly unreacted dione and the enone (1 3 ) (IR,TLC), and a very small amount of a less polar compound (TLC). Chromatography on silica using petrol 40-60°/ benzene 2/l as eluant fhrnished 6 mg. of the latter component, which had no carbonyl absorption in the

IR spectrum, and hence was not the desired compound. b) refluxing with PTSA in toluene

A solution of PTSA (500 mg.) in toluene (10 ml,) was refluxed for 5 hr. in a water separator. After addition of the dione (10), (112 mg.; 0»5 nM), in toluene (l ml.), followed by a 5 hr. reflux, the extraction procedure used in (a) yielded a brown oil (112 mg,). The product contained only the enone (l3)»

(IR, TLC), and a small amount of a less polar component (TLC), the latter probably causing a band at 1740 cm. ^ in the IR. Chromatography on silica using light petroleum 4O-6O0/ benzene mixtures gave a fraction (8 mg.) with no carbonyl absorption in the IR, and a fraction (5 nig.) with IR absorption at 1740 cm.“^, which, because of lack of material, was not further investigated. Elution with benzene, and benzene/chloroform, yielded reasonably pure enone (l^ *(83mg.). “ 52-

61 c) concentrated sulphuric acid

Concentrated sulphuric acid (l ml.) was added dropwise

to the dione (lO), (448 mg,; 2 mM), and the mixture stirred for

72 hr. in a stoppered flask. Careful addition of saturated sodium bicarbonate solution and ether extraction furnished a yellow oil

(386 mg.), containing mainly unchanged (10) and a little of the enone (13), (IR, TLC), and, as in (a) and (b), a trace of a less polar component. d) Boron trifluoride etherate 62 The method used was based on that of Corey and Nozoe.

A mixture of boron trifiuoride etherate (3 ml.), dione (i q) , (224 mg.; 1 mM), and methylene chloride (6 ml.) was heated at reflux for 4ir hr., and then allowed to cool. Addition to water, and ether extraction, yielded a yellow oil (216 mg.) containing approximately equal amounts of unchanged (lO) and the enone (13 )» (IR, TLC).

Only the enone (13), (207 mg.), was present when this material was refluxed for 9 hr. in benzene (6 ml.) with boron trifluoride etherate ( 3ml.).

1.5.5-trimethyl-9“methylenebic.vclo (5.4.0)undec-7-*ene .LffiL 76 To a suspension in dry ether of phenyl1ithium (20 mM) - prepared from bromobenzene (3*12 gm.) and lithium (278 mg.) - was added dropwise over 5 rain, with stirring in a nitrogen atmosphere a 63 suspension of triphenylmethylphosphonium bromide (7*14 gm.; 20 niM) - 53-

in dry ether (40 ml*). The resultant cloudy yellow mixturer con taining a suspension of the Wittig reagent, m© thylaneiriphenyl- phosphorane, was stirred under nitrogen for 3 hr. A solution of the enone (13), (2*06 gm.; 10 mM) in ether (30 ml*) was added dropwise under nitrogen, destroying the yellow colour, and rendering the solution much cloudier. The mixture was stirred overnight under reflux (without nitrogen), allowed to cool, filtered to remove most of the finely-divided solid, and evaporated under reduced pressure.

The yellow oily solid residue was extracted thoroughly with hot light petroleum 40-60°. The petrol extracts were filtered wider suction several times to remove colourless triphenylphosphine oxid# (PhjPO), washed with water(3x), dried and evaporated. The residual yellow oil consisted mainly of (41), with a small amount of PhjPO, and a trace of unchanged (l3)» (TLC). The products from most runs were chroma1to- graphed on silica (aluant: light petroleum 40-60°), giving good separation of (41) from its closest chromatographic contaminant,

Ph^PO. However, on one occasion, total rearrangement of (41) occurred, and later runs were purified on neutral alumina grade 3, which gave a less satisfactory but adequate separation. The diene (4L) was eluted first as a colourless volatile oil (1*402 gm.;69 $), which underwent very gradual atmospheric oxidation to the enone (13).

Despite repeated distillation, satisfactory analysis figures were not obtained. X lax* 243 mp ( £ 178O 0) -54-

IR, table 4 » HR) table 8 ; mass spectrum, table 5 *

1«5 > 5,9-tetramethylbicyclo(5.4.0)undec-7-ene (Ad)

The diene (41), (3*23 gm.; 15*83 mM) dissolved in analar ethyl acetate (15 ml.) was hydrogenated in the presence of

5$ Pd on charcoal (1*5 gm). Uptake ceased after about 45 min.

Filtration and evaporation of solvent(with minimal heating) under reduced pressure furnished (44) as a colourless, volatile oil

(3*00 gm.; 93$).

IR, table4 J NMR, table 8 ; mass spectrum, table5 •

Attempted ozonolysis of olefin (44)

10$ ozone in oxygen was passed for 15 min. through a solution of the olefin (44), (43 rag.; 0«21 mM), in ethyl acetate (5 ml.) at -70° (acetone-dry ice bath). Glacial acetic acid (2 ml.) and zinc powder (60 mg.) were then added. After being stirred at room temperature for 1 hr., the mixture was filtered, and the solvents evaporated. The residue was dissolved in ethyl acetate, and washed with saturated sodium bicarbonate solution, to separate acidic material

(6 mg.) from neutral material (29 mg.). The latter showed IR absorption at 1700 cm."*’*' (0=0), but none at 2650-2880 cm.”'*'

(C-H str. aldehyde), and consisted of a mixture of several components

(TLC). The M3R spectrum showed no aldehydic proton signal. - 3 5 -

1,5,5,9-tetraraethylbicyclo (5.4.0)undecan-7.8-diol „JA7±

S ~ t r y The hydroxylation procedure used was that of Baran.

Osmium tetroxide (3*5 gm.; 13*79 raM) was added portionwise to a solution of the olefin (44), (2*8 gm.; 13*60 nil), in dry pyridine

(50 ml,). After 3 days, the dark brown solution was stirred for

1 hr. with a solution of sodium bisulphite (6*3 gm.) in water

(105 ml.) and pyridine (70 ml.). The mixture was extracted with chloroform (5X), and the organic extracts were washed, dried, and evaporated, with the addition of dry benzene to effect azeotropic removal of pyridine. The residual diol (47) was a viscous brown oil (3*05 gin.; 93*5$) which on distillation was obtained as a colourless crystalline solid, which slowly solidified, m.p. 86*5-

88*5°.

Found (ffl'&ss anal. )240*20752 j C15H28°2 requires 240*20892

IR, table4 ; MR, table 8 ; mass spectrum, table 5 •

2-(r-formylbutyl)-2,6,6-1rimethylcycloheptanone (48) a) Attempted cleavage of the diol (47) by sodium metaperiodate

The diol in dry or aqueous methanol was stirred at room 68 temperature under nitrogen with excess sodium metaperiodate for periods up to 35 hr., but in all cases the diol was recovered unchanged. b) Cleavage of the diol (47) by lead tetraacetate

Pure diol (47), (102 mg,; 0*425 nil) in dry benzene (3 ml.) - 56- was stirreqL f*r 4 hr. with lead tetraacetate^(l50 mg.). The mixture was filtered under suction, the filtered solid washed with benzene, and the filtrate washed with water, saturated sodium bicarb

onate solution, and brine. Evaporation of the dried benzene layer afforded a neutral pale yellow oil (95 mg.) consisting mainly of (48)? with a small amount of a more polar compound (TLC).

IR, table 1 ; NMR, table 6.

Attempted preparation of 1,5,5,8-tetramethylbioyclo(4»4*l)undec-7-en- a) treatment of ketoaldehyde (48) with base A solution of the ketoaldehyde (48),(25 mg.), in methanol (5 ml.) was heated for 15 min. on the steam bath with methanolic potassium hydr#xide. Addition of brine, and ether extraction afforded a small acidic fraction, and neutral material which contained at least four components (TLC). This complex mixture was not further examined. b) p-toluenesulphonic acid in refluxing benzene on ketoaldehyde (48)

A mixture tf PTSA (7G®g») in benzene (4 ml.) was heated for two hr. in a water separator. The ketoaldehyde (48), (65 mg.), dissolved in benzene (4 ml.) was then added. The deep red solution obtained after 4-i hr. reflux was coeled, neutralised with excess solid potassium carbonate, and left overnight. The benzene solution

•btained on filtration was washed (2 X saturated NaHCO^, then 2 X brine) dried, and evaperated. The residual yellow oil contained unchanged

(48) anA three much less polar compounds. The routine IR spectrum was

* - -57- similar t® that of (48), with a weak additional shoulder at I67O cm. 1 c) treatment of ketoaldehyde (48) with concentrated sulphuric acid'^,,'^

The ketoaldehyde (48), (25 mg.), in ether (l ml.) was added to conc entrated sulphuric acid (l ml.) at 0°. The mixture was stirred at room temperature overnight, allowing the ether to evaporate. Dilution with water, followed "by ether extraction, gave a neutral fraction consisting of a complex mixture, with no starting material present. REFERENCES

1. Vogel and Roth, Angew. Chem., 196Ay J6- 145; Vogel and Boell, ibid., 1964, 26, 784,

2. Grimme, Kaufhold, Dettmeier, and Vogel, Angew. Chem., 1966, 78, 643* (Angew. Chem., Intern. Ed, Engl.* 1966, 604).

3. Knox, Velarde, and Cross, J. Amer. Chem* Soc., 1965, 82, 3727*

4* Dauben and Laug, Tetrahedron Letters, 1962* 455*

5* Dauben, Westman, and Bond, Abstracts, 141st Meeting, American Chemical Society, Washington* D.C*, March 28, 1962, p.p. 29-30.

6 . Eliel, "Stereochemistry of Carbon Compounds”, McGraw-Hill Co., 1962, New York, P.300.

7# Collins, Hobbs, and Rawson, Chem. Commun., 1967, 155*

8 . Westman and Stevens* ibid., 1965, 459*

9. Buchanan, McKillop, and Raphael, J. Chem. Soc., 1965, 833*

10. Prelog, Barman, and Zimmerman, Helv. Chim. Acta, 1949, j52> 1284.

11. Wichterle, Prochazka, and Hof man, Collection Czech. Chem. Commun., 1948* 1J5, 300, and earlier papers.

12. Stork and Landesman, J, Amer. Chem. Soc., 1956, 78, 5129* 3 13« Gritter, Sabatino, and Sacharow, Chem. Abstr., I960, 58, 3328d.

14. Schut and Liu, J. Org. Chem., 1965, 30. 2845.

15. Sands, ibid., 1963, 28, 1710.

16. Sands, ibid., 1964, 29. 2488.

17* Cope and Synerholm, J. Amer. Chem. Soc., 1950, 72, 5228.

18. Marshall and Partridge, Tetrahedron Letters, 1966, 2545*

19. Selvarajan, John, Narayanan, and Swaminathan, Tetrahedron, 1966,22,949.

20. Opitz and Mildenberger, Ann., 1961, 650, 115* 21. Loewenthal and Neuwirth, J. Org. Chem., 1967, }2, 517.

22. Bisarya and Sukh Dev, Tetrahedron Letters, 1964, 3761.

23. Hendrickson, Tetrahedron, 1959, ]_t 82.

24* Sukh Dev, Joseph, Pandey, and Bisarya, I.U.P.A.C, Symposium on Natural Products (Prague), 1962, 51*

25. Martin, Parker, Shroot, and Stewart, J. Chem. Soc., 1967, £> 101.

26. Baeyer, Ber., 1894, 2810.

27. Wallach, Ann., 1905, 339.94.

28. Barnes and Houlihan, J. Org. Chem., I96I, 26, 1609.

29. cf., Bruson and Riener, J. Amer, Chem. Soc., 1942, ,64, 2850; Prank and Pierle, ibid.. 1951» 72-4*

30. cf., Woodward, Sondheimer, Taub, Heusler, and McLamore, ibid.. 1952, 74, 4230.

31. cf., Dauben, Boswell, and Templeton, ibid., 1961, 8£, 5006.

32. cf., Zwahlen5Horton, and Fujimoto, ibid., 1957, 79, 3131

33. cf., House, Trost, Magin, Carlson, Franck, and Rasmusson, J. Org. Chem., 1965, 30, 2513.

34. Tegner, Acta Chem. Scand., 1952, _6, 702.

35. Cason, Chem. Revs., 1947, 15; Shirley, Org, Reactions, 8, 28.

36. Corey and Chaykovsky, J. Amer. Chem. Soc., 1964, 86, 1639; ibid., 1965, 81, 1345.

37* Ross and Levine, J. Org, Chem,, 1964, 22,, 2341*

38, Brown and McParlin, J. Amer. Chem. Soc., 1950, 80, 5372; Brown and Subba Rao, ibid., 1958, 80, 5377*

39* Hassner and Heathcock, J. Org. Chem., 1964, 2£, 1350, and references cited therein.

40. Martin, Ph.D. Thesis, Glasgow University, 1964* - 60 -

41* Fawcett, Chem* Bevs., 1950, J7, 219.

42. Bowden, Heilbron, Jones end Weedon, J. Chem. Soo., 1946, 39.

43* Lawson, Ph.D. Thesis, Glasgow University, 1966; Ramakrishnan, Shanmugam, and Swaminathan, Ind. J. Chem,, 1963, 1, 406.

44» cf., Fusco and Tenconi, Tetrahedron Letters, 1965, 1313*

45• Marshall and Scanio, J. Org, Chem., 1965, 30« 3019*

46, Gutsche, Org. Reactions, _8 , 364*

47* Meerwein and Bumeleit, Ber., 1928, 61, I84O; Meerwein, Bersin, and Bumeleit, ihid., 1929, 62, 999; Kohler, Tishler, Potter, and Thompson, J. Aner. Chem. Soc., 1939, 61, 1057*

48. Adamson and Kenner, J. Chem. Soc,, 1939, 181*

49• Maercker, Org. Reactions, 14, 270; Trippett, Quart. Revs., 1963, 11, 406; Trippett, Advan. Org. Chem., i960, 83.

50. Levine, J. Amer. Chen. Soc., 1958, 80, 6150; ¥ittig and Knauss, Angew. Chem., 1959, 71, 127; Vfittig and Schlosser, Ber., 1961, 2i, 1373; Wittig, Boll, and Kruck, ibid., 1962, <£, 2514.

51. Bachman and Struve, Org. Reactions, JL, 38.

52. Bergmann, Ginsburg, and Pappo, ibid., 10, 205; Stewart and Westberg, J. Org. Chem., 1965, 30, 1951, and references cited therein.

53* Box and Yoder, J. Amer. Chem. Soc., 1921, 43, 2097*

54* Perkin, J. Chem. Soc., 1899, 75, 921; Jones and Scott, J. Amer, Chem. Soc., 1922, 44., 407.

55* Pavia, Vylde, Wylde, and Amal, Bull. Soc, Chim. Fr., 1965, 2709*

56. Royals and Robinson, J. Amer. Chem. Soc,, 1956, J8, 4161, and references therein.

57* Bauben and Ashcraft, ibid., 1963, j8£» 3673.

58. Finnegan and Bachman, J, Org. Chem., 1965, 30, 4145* 59* Bhacca and Willi?jns , Applications of 3®tR Specrtrosoopy in Organic Chemistry, Holden-Day, 1964* p. 77*

60. Baisted and Whitehurst, J, Chem. Soc., 1961, 4089*

61. Dauben and McFarland, J. Amer* Chem. Soc., i960, 82, 4245<*.

62. Corey and Nozoe, ibid., 1963, 8£, 3527*

63* Wittig and SchiJllkopf t Ber., 1934» 8J, 1310? Org* Syntheses, 66.

64* Sondheimer and Meehoulam, J. Amer* Chem. Soc*, 1957» 79* 5029*

65* Biichi and MacLeod, £1962, 84* 3205.

66* De Graw and Bonner, Tetrahedron, 1962, 18, 1311*

67* Barnn, J. Org. Chem*, I960, 25, 257*

68* Jackson, Org. Reactions, 2 f 341*

69* Criegee, Newer Methods of Preparative Organic Chemistry, Wiley - Interscience, New York, 1948, pp. 1-20.

7 0, Part II, reference 49, p. 75*

71 # Murray, Parker, Raphael, and Jhaveri, Tetrahedron, 1962, 18, 55*

72* Campbell, Islam, end Raphael, J. Chem, Soc., 1956, 4096.

73* Org. Syntheses, Coll. Vol. II, 464*

74* ibid., Coll. Vol. II, 278,

75* Maercker, Org. Reactions, 14* 399*

76, Org. Syntheses, Coll. Vol. Ill, 757* P A R T II

STUDIES IN THE 3-AZABIGICLO (3*3*1) HONARE SYSTEM iNTuniiumnw

The 3-azabicyclo (3.3*1) nonane system (l) constitutes the

nitrogen - containing part of the ring skeleton of the diterpene

alkaloids^, e.g., atisine (2), veatchine (3) and cuauchichicine (4)*

It also exhibits structural similarities to other azabicyclo (3*3*1) nonane derivatives, some of which occur naturally, e.g. pseudopelletierine (5)*

Theoretically, the 3-nzabicyclo (3*3*1) nonane system (l) can be synthesised in two ways, viz, condensation of the nitrogen- containing bridge to a cyclohexane ring in the 1,3-positions (6), and construction of

the carbocydicring by attachment of a three-carbon unit to a 3*5-

disubsiitutedpiperidine (7).

In practice, however, route (6) offers wider preparative scope.

3—azaMcyclo (3*5*1) nonane (1,R=H) was first obtained in 1932 by Komppa^, by reduction in low yield of hexahydroisophthalimide at the lead cathode,

file ii'ids was prepared by distillation of the Harmonium salt of cis/trams hexahydroisopfctbalicacid (8). Since only the cis form would be ejected

to undergo cyclisation, it appears that under the reaction condition the

trfiifis fern is converted into the cis form.

This method has found most application in the synthesisof

3-sabstitutedderivatives of (l), some of which possess hypotensive activity.

Rossi e t a l? prepared 1 (.141), which on N-alkylationformed 'the compounds (9) *

Treatment of (8) with methylomine, followed by reduction of the resultant

iwide, 1 (iMMoP, which was converted into its 2-diaethylaminomethyl derivative^. The corresponding 5~derivative was also synthesised^-.

Condensation of 1 (R=H) with acrylonitrile and simple epoxides"*, and reaction of the anhydride of (8) with dialkylaminoalkylamines^1*^ yielded a variety of compounds of type (10). Rice and Grogan*^

synthesised a number of compounds of type (ll) by the latter route.

Concomitant work by Schneider and Gotz®»9 afforded

l(R=Me,Et,CH2Ph) by reacting (8) with primary amines, followed by reduction of the imide in high yield with lithium aluminium hydride

(the reducing agent employed in all the more recent reports). Similarly,

Leonard et al.^ prepared l(R=CH2® 2® ) from 2-aninoe thanol and either

the dimethyl ester of (8), or the ditosylate (12).

An important method of accomplishing route (6) makes use of

the Mannich reaction. Heterocyclic bicyclic compounds of the pydine (13)

and bispidine*^ (14) type have been made in this way from pyridone - I and pyrone-oc,oc - diesters, methylamine and formaldehyde by Mannich et al.

In 1957» Weatherbee ejt al.^-5 attempted unsuccessfully to prepare

isopseudopelletierine (15) by condensation of cyclohexanone with

formaldehyde and methylamine, only the diketoamine (16) being isolated.

In 1959, however, Blicke and M c C a r t y ^ , using the more activated

cyclohexanone - 2,6- diesters, prepared the 1,5-diesters (17) in good

yield. The trihydrochloride of (18) was similarly obtained1^ from

the corresponding diaminocyclohexanone^5, At the same time, the

" 16 9 diester (17a) was synthesised by Schneider and Gotz ”, who hydrolysed -64-

and, surprisingly, de carboxyl at ed it to (l5)r Wolff-Kishner reduction of which afforded l(R=Me), identical to the compound made by reduction of the corresponding imide. In 1962, House et al.^ succeeded in synthesising (15)> albeit in low yield, by Mannich condensation of cyclohexanone. Another product was found to be the diketoamine (16).

In their synthesis of isopseudogranatoline (19)» Rossi and

Butta^-® prepared the derivatives (20a-d) of (15) by means of the Mannich reaction on 2-substituted and 2,6-disubstituted cyclohexanones. Each product on treatment with acid furnished (15), which was reduced by sodium borohydride and oxidised by potassium permanganate to (19)*

The similar bicyclic aminoketones (20d-g) were synthesised by 19,20 Shimizu et. al. during investigations on a total synthesis

(vide infra) of the skeleton of the atisine series of alkaloids, 22 The preparation of the derivatives (22) of (15) by Becker is the only example of route (7). The 1,2,3-trisubstituted 4-piperidones

(21) on treatment with methyl vinyl ketone in the presence of triethylamine afforded the bridged azabicyclic ketols (22) in high yield. This is a synthetic method commonly used to prepare carbocyclic bicyclic systems. Treatment of (22) with sodium methoxide effected retroaldolisation and re-cyclisation to the decahydroisoquinolines (23).

There have been two reports of syntheses of a partial skeleton of the diterpene alkaloids (24). Mercuric acetate oxidation of l(R=CH2Qa2CH) (vide supra) yielded the tricyclic oxazolidine (25), used by Leonard et al.^~® as a simple model of the A,E,F—ring system of -65-

the hexacyclic diterpenoid alkaloids. In (25), there was no steric driving force to the rearrangement of the oxazolidine ring which occurs ii. these alkaloids, e.g., atisine (2) and isoatisine (26).

The model (25) served to define the stability towards isomerisation, and the base strength and other properties of the tricyclic ring moiety in (24) when freed from the steric influence of the remaining rings B,C and D.

The starting material in a synthetic approach to diterpenoid alkaloids by Shimizu et al.^9>2l was (20f) (vide supra). A cjs/trans mixture of the tetrahydrophenanthrene derivatives (27) was eventually obtained (Scheme l). Since atisine-type alkaloids possess an A/B-trans conformation, only the trans isomer of (27) would be of use in further elaboration to the diterpene alkaloids.

Stereochemical studies have been carried out on derivatives

Of (1) using chemical and physico-chemical techniques. House et al.^ concluded that the absence of irregularities in the infra-red and ultra-violet spectra of (15) suggested that there was no appreciable interaction between the amine and carbonyl functions. Also, a series of three peaks in the infra-red spectrum of (15) in the 2600-2800 cm.”^ region indicated*^ that the favoured conformation had at least two hydrogen atoms trans and coplanar with the unshared electron pair on the nitrogen atom. The N-methyl group must, therefore, be equatorial as in (28). Dipole moment and pK^ data-*-? were also in accord with (28).

Studies of the decarboxylation2^, reduction2^, reaction -66-

with organometallic reagents^1, U—alkylat.ion^^ and aolvolysie^ of (15) or its derivatives added further evidence for the predominance

of the chair-chair conformation.

Pumphrey and Robinson*^ deduced from proton magnetic resonance studies on 7-t-butyl derivatives (29) of (l) that both the carbocyclic and the heterocyclic rings were in the chair conformation

(30). The configuration of the nitrogen atom was established by the 17 presence of infra-red absorption similar to that observed by House •

In the two-chair conformation, this was possible only if the unshared electron pair was in the extremely hindered endo position, proving

that the unshared electron pair is smaller than the hydrogen atom bonded to nitrogen.

Nobler and Dunitz^ reported an X-ray structure analysis of the hydrobromide of 1(R«H) and stated that the cation had a chair-chair

conformation in which the bond angles deviated somewhat from their ideal values, presumably because of steric repulsion between a pair of non bonded hydrogen atoms. Nuclear magnetic resonance studies by

McKenna et al?1 indicated that for the hydrochloride of l(R»Me) in

solution, the main configurational isomer had the chair-chair conformation

(31). The corresponding methiodide was, however, apparently almost

exclusively boat-chair (32).

Fairly recently another synthetic route involving the Mannich

reaction has appeared. Investigations on the reaction of nitroaromatic -67-

compounds with sodium borohydride by Severin et_ al* included

the postulation that reduction of the metadinitrobenzenee (33) by

sodium borohydride formed the di-sodium salt (34)^, which on

acidification afforded (35)^« Treatment of (34) with formaldehyde

and piperidine gave (36)^» the structures being proved by conversion of 36(R=Me) to the toluidine (37)» which was prepared by an independent synthesis. Mannich condensation of (34) with formaldehyde and methylamine produced in reasonable yield the l,5-dinitro-3-methyl-3-azabicyclo (3»3«l) nonenes (38)55. The structures (38) were assigned mainly by analogy;

the actual evidence presented was meagre, viz, analytical figures, ultra-violet absorption below 210 m^, unchanged in alkaline solution, and

(in two cases) approximate molecular weight determinations.

Similar products (39) and (40) were later obtained in low yields from (34) a^d the corresponding salt of 3»5-dinitroanisole respectively, and a Grignard reagent (R*MgBr), formaldehyde and methylamine-^. Apart from the proof mentioned above, idle compounds exhibited absorption in the infra-red at 1540cm“^ (nitro group) and

since no silver chloride was formed by treatment of 39 (R=Cl) with alcoholic silver nitrate, the chlorine atom was stated to be vinylic.

More complete spectroscopic proof was included in a later study, in which 2 ,4-dinitrophenol in the presence of sodium ethoxide reacted with

acetone to form the salt (41)* which gave the bicyclic compound (42)

on reaction with methylamine and formaldehyde37. Cyclohexanone -68-

■57 behaved analogously to form (44) via (43) • Sodium borohydride reduction of (41) and (43) to (45) and (46) respectively, followed immediately by Mannich condensation, yielded not the expected products,but the tricyclic compounds (47) and (4s).37

Previously mentioned examples of Mannich reactions used in route (6) involved activation by carbethoxy groups of the sites of

Mannich condensation, whereas Severin’s novel procedure made use of activation by nitro groups in salts such as (34)* These salts promised to be of wide synthetic utility in the preparation of bridged bicyclic systems with a nitrogen atom at each bridgehead. It was therefore decided to seek proof of the structures (38) using rigorous chemical and spectroscopic means, and to examine thoroughly the scope

of the reaction sequence leading to them. h o 2c TsOH2C\/CH 2OTs

11 R — H ,CH3, n~Bu, (CH2)nNMe2 ,(CH2)nNEt2,n=2,3

? h 3 r ( K C> R R 1N K l

V y CH CH C H

13 K 15 R ^ H 17a R = C 0 2Et 17b R = C 0 2CH 3

0 H OH CH h 16 19

R R; 20 a C 0 2Et R R b C O N H P h c H C O N H P h V d H C 0 2Et CH e CH 3 C 0 2Et f CH 3 C 6H AO C H 3-b CH 3 C 6H AO C H 3- m 3 R 0 2C 0 H 2 A r 0 OH

c o 2r CH- Ac . ^ C H 2 A r ^ c o 2r C H2 A r A c Y OH A c 21 22 23 R = C H 3 A r = Ph t-Bu P h CHo C g H ^ O C H g- !0 CH CgH 3 ( 0 0 H 3 )2 “ 3,4

^ c ^ O C H

OH 3 N SCHEME 1

C 6H aO C H 3-|d c h 3 -C=COEt 0 Li C=COEt OH H

C H = C H O E t

LAH Pd/Cf A A C H 2CH O H2 ^ "'CHCHO

OCH

A / P P A j 27

29 30 R = H, C H 3 3 R( f ^ l o 9 n ^ n o 2

31 32 33

R f ^ 2 N a R O o N NO.

34 35 R = C H 3, Br,CH=CHPh,C02H R = H 36 R= as above

R = C H 2-r^

ch 3 h2n^ c h 2-hi O o N NO.

r ch 3 37 38 R = H , C H 3,Cl,CH=CHPh OCH3 Et 2 02N Et NO

VCH.

39 AO AT R = C H 2COCH 3 R -H ^R = CH3,Et,n-Bu 43 R = R=Cl,,R=Et 0

45 R = C H 2CHCH 3 OH

46 R = HO

° 2 IK n ^ C H 3

42 R = C H 2COCH 3

44 R = 47 -69-

DISCUSSION

It was firstly necessary to repeat the proo-odure of 35 Severin et al # for the preparations of the known 1,5-dinitro-3- methyl-3-azabicyclo (3«3*l) non-6-enes (la), (lb) and(lc), in order that their structures might be confirmed by rigorous chemical and spectroscopic methods. Hence were obtained crystalline compounds, with melting points in agreement with those reported. Moreover, their mass spectra indicated the expected molecular weights of 227, 241 and

261 (263) for (la), (lb) and (lc) respectively, and also contained peaks which confirmed the presence in each compound of two nitro groups (Table II). Further evidence for the proioosed structures was available from their IR spectra (Table l), which exhibited bands at ca. 2800 (N-methyl), 1550 (asym. nitro) and 1340 (sym. nitro) on.”-5-

In the case of (lc), doublets for the nitro group frequencies at

1562, 1552 and 1368, 1343 cm.""'5' could be attributed to interaction of the chlorine atom with the adjacent nitro group. A peak due to the double bond of (lc) was also observed at 1555 cm.”-5- However, the HMR spectra (Tabled) of (la), (lb) and (lc) presented somewhat conflicting evidence. Althou^i those of (lb) and (lc) appeared to be in accord with the proposed structures, that of (la) contained a sharp singlet in the olefinic proton region at 3*^9 (2H), and a broad singlet in the allylic proton region at 7*20. This observation cast doubt on structure (la), and prompted an extension of Severin’s procedure to the syntheses of a -70-

series of compounds of "types (l) and (2), so th-ai a. ctoir© detailed

examination of uheir Nl-ffi spectra could be made.

2,4-Dinitroanisole yielded after sodium borohydride reduction and Mannich reaction, 30fo of crystalline material, analysing for c10h15n3°5» and with the correct molecular weight of 257 (mass spectrum).

Its IEt and KMR. spectra appeared to confirm the presence of the bicyclic vinyl ether (id). The former exhibited the peaks expected for N-methyl and nitro groups, with additional absorption at 1680 cm.' -1 (c=c).

The latter included a triplet at 5*13 (1H> J=4c/s* olefinic proton), a doublet at 7*15 (2H, J=4c/s, allylic proton), a singlet at 6*45

(3H, methoxyl protons), and a singlet at 7*58 (3H, NCH^). In contrast with this spectroscopic data, the chemical behaviour of this compound was not that expected of a vinyl ether, since only starting material was recovered after heating with dilute acid. With concentrated acid, no product at all was obtained. A probable explanation of the latter result was that the unstable intermediate ketone (3) decomposed to the water-

soluble aminoacid (4)*

From the appropriate m-dinitrobenzenes were prepared (le),

(if), (2a), (2b) and (2c). Correlation between two compounds of type

(2) was provided by esterifioation of the acid (2a) to yield crystalline material identical to the ester (2fc). Considerable variations in yield were observed between compounds of types (l) and (2). The 7—substituted

compounds (2) were available in much higher yields than the 6-substituted - 71-

analogues (l), e.g. 7-C02St,7-C02Me, 98$; 6-C02Et, 4$; 7-OMe

(vide infra), 66$; 6—OMe, 3Q$« Stability was not however related to structure (l) or (2). On exposure to light, (la) and (id) were rapidly yellowed, whereas (2b), (2c) and (2d, vide infra) remained unaffected even after longer exposure.

Only the product from 2,4-dinitrophenol was not of the expected type. Hot ethanolic extraction of the crude reaction product, followed by recrystallisation, gave ca. 1$ of colourless needles analysing for C9H16N4O4 . The mass spectral molecular weight of 244 was in accord with the analytical figures. The presence of two nitro groups was shown by peaks at 198 and 152 n/e, corresponding to losses of 46 (N02) and 92 (2NO2) m/e. The extra nitrogen atom implied by the analytical data was confirmed by the IR and NMR spectra.

Peaks at 2812 and 2798 cm.”'5' in the IR were attributed to two N-methyl groups, and those at 1551 and 1353 cm.”'5' to nitro groups. The NME spectrum contained a singlet at 7*84 (8H, two NCSIj), and an AB quartet at 7*13 (j=llc/s, bridge protons) with an overlapping signal at 7*34

(1CK in all, 4 NCH2) • These data were fully consistent with the bispidine structure (5)> the probable mode of formation of which is outlined in Scheme A. The instability of a molecule possessing a carbonyl group adjacent to a nitro group was reported by Severin et aL, who found that the sodium salt (6) of picric acid, on reduction with sodium borohydride, followed by acidification, gave the acyclic -72-

compound (7) (Scheme B). Similar reactions are known in the literature, the formation of 1,5-dinitropentane by acid 43 treatment of the mono-potassium salt (8) of 2,5-dini trocyclohexanone.

A detailed examination of the spectra of these compounds was now undertaken.

The IR spectra (Table 1) of both types (l) and (2) showed

N-methyl absorption at 2790-2800 cm.”-*- Absorption due to the double bond fell in the range 1655-1865 cm.”'5', except when R=0Ke (1680 and

I669 cm. ^ for (id) and (2e) respectively). Some compounds of type

(l) exhibited a doublet (Av 6-14 cm.”'5") for the asymmetric nitro group frequency in the 1550 cm.”'5' region, indicating interaction of the group R (Cl, OMe, C02Bt, C02 H) with the adjacent nitro group. The doublet observed in the ester carbonyl region for (if), (R=C02 Et) was also explicable on this basis. In contrast, compounds of type (2) exhibited only single peaks in the 1550 cm.”'5' region. The symmetric nitro group frequency was observed as either a single band at 1340-

1350 cm.”-5-, or a doublet at ca, 1365 and 1345 cm.”-5-, with no apparent correlation with structure type (l) or (2). Bohlmann bands in the

2600-2800 cm.”1 region2^, which indicate that the favoured conformation has at least two hydrogen atoms trans and coplanar with the unshared electron pair on the nitrogen atom, were not observed.

All the mass spectra (Tables H andl2) contained peaks corresponding to losses of 48 (N02), 47 (HN02), 92 (NO2 from P-46) and 93 (HNO2 from P—46) ra/e, and a peak at 91 m/e, probably due to the

ion C7H 7 (compounds without the double bond (Tabled) did not

exhibit this peak). In most cases, losses of 30 (NO) and 60

(NO from P-30) were also observed. The (P-l) peak (loss of a 44 hydrogen atom), which has been reported as being characteristic

of N-methyl piperidines, was not present. Mass spectrometry hence

confirmed the presence of two nitro groups, as well as identifying the

substituent R.

The NMR spectra of compounds of type (l), (R$ Ii) exhibited

a triplet (J=4c/s) in the olefinic proton region, and a doublet at ca. 7*1 (j=4c/s) for the allylic protons. In the case of (ll), these signals were observed as multiplets with half-band widths of 7-8c/s

(Table 14). The 7-substituted compounds (2) showed broad singlets for the olefinic and allylic protons, each with the same half-band widths (Table15).

As mentioned above, the olefinic protons of (la) resonated as a sharp singlet at 3*89, and the allylic protons as a broad singlet at

7*20, which raised doubts as to the correctness of the proposed structure.

Similar signals were present in the spectra of (34) and the picrate of

(35)» which were prepared later (Tabled). In trifluoroacetic acid, the olefinic protons of (la) and (34), (Table 17), gave rise to doublets at

3*75 and 3*70 (each J=6c/s) respectively. In benzene (Table 18), finer splitting was observed for (la). The C6 proton gave a doublet at 4*43

(J=2.3c/s), and the Cy proton a triplet at 4*87 (J=3c/s), each superimposed -74-

on a multiplet. It thus appeared tho-t in CDCl^, the sharp singlet observed for the olefinic protons on iho unsubstituted double bond was due to accidental equivalence.

The singlet due to the N-methyl protons appeared in the range 7*53 - 7*82 for compounds of types (l) and (2).

Difficulty was experienced initially in assigning the absorptions due to the four ITCH? protons. It was considered that by comparing the

NMR spectra with those of the corresponding picrates, the NCH2 proton 45 signals could be identified by their known downfield shift in the picrate relative to the parent amine. The picrates of (la)-(ld),

(if), (2b) and (2c) were prepared, but after several recrystallisations, these rather insoluble derivatives still exhibited a n.p, range of ca.

10°. However, the picrate of (35) > which had a n.p. of 134-144°, was lat er found to analyse correctly.

With the exceptions of those of (lc) and (2b), these picrates were almost totally insoluble in CDClj, The two soluble picrates however gave signals identical to those of the parent base, in addition to a singlet at 0*80 (2H, picrate anion aromatic H),

This was a surprising result, since work in this department had revealed considerable shifts (ca. 0*7 ppn) for the crfproton signals in the picrates 45 of Mannich bases. It was therefore decided to initiate a study of the NMR spectra of simple amines and their picrates, in order to elucidate the reasons for the observed shifts. A common solvent (CDCI3) -75-

was used throughout in the expectation that any sol vent shift would be eliminated, and a more regular displacement pattern might emerge.

In fact, the results (Tablel9) showed a pattern very similar to 46 that reported. Tertiary amines showed a larger shift than secondary amines. In the former, A (KCK2) was ca. 0*75 ppn., in the

latter, 0*48-0*54 ppm. Displacements of |BCH signals, where they could be identified with certainty, were smaller (ca. 0*3 ppm). In one case (Et^N, Table20), the hydrochloride was found to show similar shifts to the picrate, suggesting that the aromatic ring played no part in the deshielding, which must be entirely due to N-protonation.

¥ith this amine, similar shifts were observed in D20 as in CDCI3 (Table20)*

Since the picrates of (lc) and (2b) did not show these shifts, it was considered possible that in CDCl^ solution dissociation might have taken place, resulting in a mixture of picric acid, and the tertiary amine. Evidence for this came from the M'ffi signal due to the picrate anion aromatic protons. In the simple amine picrates, this signal occurred at 1 *1 , whereas in the supposed picrates of (lc) and (2b), singlets at 0 *80 were observed. Since for picric acid, the aromatic signal occurred at 0 *80f NMR spectroscopy seemed to confirm the presence of picric acid.

In contrast, material precipitated from CDCI3 solutions of the picrates of (lc) and (2b) possessed melting points in the same ranges as the picrates. These melting points were far higher than that of -76

picric acid. Either complex absorption or a pair of broad singlets at 7*2 - 7*7 was eventually assigned to the NCE2 protons.

The geminal bridge protons of the compounds (1), (R^H) , appeared as a pair of broad singlets at ca. 6.6 and 6*8 (j ca. 12c/s).

The compounds(2) and (la) gave apparent triplets at ca. 6*8 (j ca. llc/s), presumably due to overlapping of the two inner peaks of the IB quartet.

In two saturated compounds later prepared, an apparent doublet

(j=12c/s), each signal slightly split, was present. This was considered to arise by interchanging of the two inner peaks (Table 22). This signal, shifted considerably downfield by the two nitro groups, varied little in position in all the compounds studied,

The preparation of 1,3,5-trinitrocyclohexane (9) by sodium borohydride reduction of sym-trinitrobenzene (TNB) has been 32 reported by Severin ejt al. Since comparison of its spectra with those of compounds (l) and (2) would be of interest, the published procedure was repeated, yielding a highly unstable red oil, consisting mainly of one compound (of, (9) was reported as having m.p. 125°) •

This material was probably crude (9)» since absorption in the IR occurred at 1§63 cm.“1 (NO2 asym. str.). Such compounds are known to 47 be unstable, e.g., it was observed by Nielsen that 1,3-dinitro- cyclohexane was unstable under basic conditions.

Similar reduction of TNB, followed by treatment with formaldehyde and methyl amine, did not afford the bi cyclic trinitro compound (10) • The basic fraction of the product exhibited nitror

N-methyl, and also weak hydroxyl absorption in the IR, and although shown by TLC to consist mainly of one compound, it could neither be induced to solidify, nor could a solid picrate be prepared.

It had therefore been demonstrated that the reduction -

Mannieh sequence was applicable to a wide range of substituted n- dinitrobenzene s. p-Dinitrobenzene, by analogous behaviour, would be expected to form the 3-azabicyclo (j.2.2) nonene (ll), However, no violet coloration (characteristic of the m-dinitrobenzenes) was formed on addition of sodium borohydride, and after treatment with formaldehyde and methylamine, the small basic fraction formed was found to contain none of the desired compound (ll)# Although similar treatment of o-dinitrobenzene was not investigated, it seems feasible to state that this reaction sequence is limited solely to meta- substituted dinitroaromatic compounds.

Although further extensions in the scope of this sequence were cs&rried out (p.87) it is pertinent at this point to detail the chemical studies on compounds of types (l) and (2) which served to provide added proof of structure#

Hydrogenation of (la) with IQffo palladium cn charcoal in ethyl acetate furnished yYfo of the saturated compound (12), after chromatographic separation from more polar products* The mass spectral molecular weight of 229,and the lack of olefinic proton resonance in the -78-

NMR spectrum, and of double bond absorption in the IR spectrum, confirmed the presence of one double bond in (la).

Efforts were next directed at the degradation of (la) by oxidative cleavage of the cyclohexene ring to form a 3 ,5-dinitro-

N-methyl piperidine. Accordingly, an aqueous suspension of (la) 48 and selenium dioxide vfeueheated at reflux for 18 hours, but, despite various extraction procedures, only about 2Cffo of material was recoverable. The main component, isolated in ca. \Ofo overall yield, was a colourless crystalline compound, n.p. 68-9°• A singlet at

0»37 (lH, formyl proton) and a doublet at 3*62 (2H, J=6*5c/s, olefinic protons on a oio double bond) in the HKR spectrum indicated the presence of an unsaturated aldehyde. An AB quartet at

6*10 (2H, J«12c/s) showed that the bridge methylene had remained intact. No signal was observed in the 7*6 region (NCH3). Remaining peaks at 6*89 (singlet, 4H) and 7*13 (broad singlet, 4H) were not assigned. IR spectroscopy confirmed the above observations, viz., no N-methyl absorption, and peaks at 1702 (^JJ-unsaturated aldehyde

C=0) and 1551 cm."3* (asym. NO2 str.). TTV absorption occurred at 49 ^ EtOH 227 mu (e3600), (cf, crotonaldehyde, Xmax 2^7 ku» el7900). max / 1 The material analysed for C9H14.N2O4 (mol. weight 214) > in disagreement with the mass spectral molecular weight of 242. However, no peaks could be accounted for by loss of NO2 or NO, and the spectra of two later products, although identical as far as 170 m/e, had different -79-

parent ions. Although soluble in dilute hydrochloric acid, it did not form a picrate. Positive Fehling1 s and silver mirror tests were however observed. Two structures, neither of which was compatible with the foregoing conflicting evidence, were briefly considered.

The aldehyde (13), (243)» the formation of which is rationalised in Scheme >C, however possessed a N-methyl group. The amide (14)» C9H11N3O5 (241), arising by oxidation of the N-methyl group, would be expected to have a NMR singlet at ca. 2*0 (cf,

H00Niyfe2, 2*16), and an IR band in the range I67O-I63O cm.~^

(tertiary amide 0=0). A satisfactory structure for this product was not obtained.

It was decided to employ a less random method of cleaving the double bond of (la). Hydroxylation of (la) with osmium tetroxide by the procedure of Baxax!?^ yielded 93% of the 6,7-cis-diol (15)»

C9H15N3O6. The mass spectrum showed the correct molecular weight of 261, with peaks corresponding to losses of 17(0H), 18(H20),

46(N02), 92(2N02) and 93(N02 and HN02) m/e. The loss of 17 m/e may have arisen by the breakdown process depicted in Scheme D. The diol exhibited the expected IR absorption (hydroxyl, N-methyl and nitro).

Due to insolubility in deuterochloroform, the NEffi spectrum was run in trifluoroacetic acid, but apart from a singlet at 6*71 (NCH3), assignments could not be made*

Treatment of the diol (15) in methanol with sodium metaperiodate -30-

in a nitrogen atmosphere, afforded g. "browo semi-solid oil * wiiioli showed hydroxyl, N-methyl, nitro and carbonyl (1720 cm*“^) absorption in the IR, Careful recrystallisation yielded an unstable pale-brown crystalline solid with a mass spectral molecular weight °£ 231.

Significant peaks in the mass spectrum were at 230 (P-l) and

214 m/e (P-17> loss of OH), (the latter probably arising by a process similar to that in Scheme D), as well as those indicating the presence of two nitro groups. Absorption in the IR (CHCI3) occurred at 3595

(sharp peak, free hydroxyl), 2805 (N-methyl), 1549 (asym. N02) and

1371 (sym, N02) cm, The carbonyl region was transparent. These data were compatible with the 3-azabicyclo (3.2.1) octane alcohol

(16), formed by loss of formic acid from the intermediate dialdehyde

(17), followed by cyclisation (Scheme E). The IR carbonyl absorption in the crude product could then be attributed to the dialdehyde (17),

M R assignments were made as follows:-

Multiplet at 5*5 (lH, half-band width ca. 12c/s, C5 carbinyl proton); an apparent triplet at 6*73 (2H, J=10c/s, bridge protons); 7*42, ca.

7*6 (NCH2), ^ singlet at 7*59 (ICE in all, NCH3)• The alcohol (16) was too unstable to be prepared analytically pure, and on standing was decomposed to amorphous material insoluble in chloroform. Since the odour of nitrous acid was detectable, the decomposition pattern was probably as outlined in Scheme P.

The hydroxyl group of (16) was shown to be extremely hindered -81-

by the recovery of unchanged (l6) after treatment vith excess Jones reagent. At this point, it was planned to convert the tosylate of (16), 51 by treatment with refluxing collidine, to the olefin (18), oxidation of which would afford the desired 3,5-dinitro-N-methylpiperidine.

In the event, exposure of (16) for 16 days to tosyl chloride in pyridine gave a product, the IR spectrum of which showed only very weak tosylate peaks at 1600 cm.“^ and in the fingerprint region. TLC analysis detected only starting alcohol.

Conclusive proof of the structure of the alcohol (16), and correlation between the compounds (la) and (lb), was afforded by similar hydroxylation of the olefin (lb). The resultant crystalline diol exhibited IR and M R absorption consistent with structure (19)«

In particular, the M R multiplet at 5*45 (lH), due to the Cj carbinyl 52 proton, had a half-band width of 18c/s, demonstrating that -this proton was axial, i.e., the Cj hydroxyl group must be equatorial (exo ).

Reaction of the diol (19) with sodium metaperiodate gave, after chromatographic separation from unchanged diol, material identical to the alcohol (16) obtained from the diol (l5)» In this case, the intermediate ketoaldehyde (20) must have lost acetic acid prior to cyclisation to form (16).

Since the presence of the bridgehead nitro groups appeared to result in decreased stability, and undesired decompositions, it was decided to attempt their removal by deamination of the corresponding -82-

triamine• The dinitro conpound (la) was therefore treated with 53 excess lithium aluminium hydride in refluxing THF, affording 6 % of the triomine (21) as a yellow oil which darkened on standing.

Owing to extensive tailing on TLC chromatograms, even in amine buffer solvents, the purity of this material could not be estimated.

It however gave rise to IR absorption at 3350 (broad band, N-H str.),

5050 (*C-H str.), 2800 (NCH3) and 1610 cm.-1 (N-H def.). No nitro group absorption was present. Distillation proceeded only with large loss, and attempts at preparing solid derivatives proved unsuccessful* 54 A deamination procedure used by Berson and Ben-Efraim was applied in an attempt to convert crude (21) to the diacetate (22).

Overnight treatment of (21) with acetic acid and sodium nitrite gave a low yield of a dark yellow gum, which exhibited similar TLC behaviour and IR absorption to the starting material. An additional peak in the

IR at 1730 cm,""'*' (acetate 0=0) was however present. Several changes in reaction conditions led to similar products. 55 An alternative method due to Stetter and Goebel was also employed. It involved sodium nitrite and aqueous acetic acid, and was used by them to convert an amine to an alcohol in high yield. Under these conditions, crude (2l) however furnished material similar to that obtained before.

No 3“&zabicyclo (3.3*1) nonan-7-ones have been reported in the literature, and since it was of interest to examine the nature of any -83-

interaction between the amine and carbonyl functions in this bicyclic

system, it was cbcided to attempt the synthesis of the aninoketone

(23)* Also planned was the preparation of the dibenzylidene derivative of (23), which by ozonolysis would afford the desired dinitropiperidine«

The first sequence aimed at (23) consisted of hydroboration of (la), followed by oxidation* It was however anticipated that 56 hydroboration of the unsubstituted double bond of (la) would result in the formation of a mixture of the alcohols (24) and (25)* In contrast, the substituted olefin (lb) would be expected to yield solely the exo alcohol (26), Hence, the preferred starting material would be (lb). The alcohol (26) was accordingly prepared in low yield, after chromatographic separation from unreacted (lb). Only an IR band at

3623 cm."*^ due to a free hydroxyl group, was present in this region.

The M R spectrum also pointed to an exo conformation for the 7- hydroxyl group, since the broad multiplet at 5*45 (1H, C7 carbinyl proton) had a half-band width of 20c/s (diaxial coupling of C7 proton with C5 and Cs protons).

Oxidation of the alcohol (26) with Jones reagent afforded a -1 mixture of three compounds, having IR absorption at 1720 and 1680 cm.

The low yield on hydroboration, and the formation of a mixture on oxidation, made it obvious that this route to (23) was unsuitable.

TJnsaturated amides are known to undergo the Hofmann - 84 -

rearrangement to yield ketones. In pr.urti-oulao;, Bell and Aroher^ have reported the preparation of 2-tropinone (27) by treatment of 41 anhydroeegonine amide (28) with cold aqueous sodium hypochlorite in methanol, followed by acid hydrolysis. The unsaturated amide

(2d) was therefore prepared from 3 ,5-dinitrobenzamide in good yield in the usual way. This stable crystalline compound analysed for

C1QH14N4O5 , and showed the correct molecular weight of 270 from the mass spectrum. IR (nujol) bands at 3458 (free N-H str.), 3170

(bonded N-H str.), 1696 (amide I C=*0), 1618 cn.~^ (amide II 0*0) and 1659 cm. ^ (C=C str.) confirmed the presence of an unsaturated amide. In the NMR spectrum were observed a broad signal at 2*4

(2H, amide protons), a broad singlet at 2»84 (lH, olefinic proton), and a, singlet at 6*65 (3H, N-methyl protons).

Low solubility in methanol of the amide (2d) necessitated the addition of a considerable amount of TEIF prior to the addition of the sodium hypochlorite solution. After treatment with dilute hydrochloric acid, a rather laborious procedure to remove unchanged amide furnished a clear red oil, a mixture of five compounds. The desired ketone (23) was isolated from the least polar band of the preparative thin layer chromatogram in an overall yield of 12% (based on consumed amide). The colourless needles analysed for C9H13N3O5 , and exhibited the following NMR signals:- a triplet at 6*85 (2H, bridge protons), a singlet at 7*04 (4H, C6 and C8 protons), singlets at -85-

7*21 and 7*40 (4H, NCH2), and a singlet at 7.59 (3H, RCH3). ^nder

conditions in which cyclohexanone formed a bis-benzylidene derivative*

the aminoketone (23) however yielded no product. Its IR spectrum had V q^q4 1734cm.and Y 1727 cm.~^ These unexpectedly high values prompted comparison with those of the ketones listed in

Table 5 • The observed Y for (23) of 1734 cm.“^ is seen to be

the highest of all the frequencies listed, although that of the dione

(29), (1729 and the highest value reported for the aaninoketone

(30) (1733 cm.”^), are of the same order.

Leonard found that in cyclic aminoketones such as (3l)»

the carbonyl stretching frequency was decreased by transannular association of the carbonyl and amine functions (Scheme G), The IR spectra of the perchlorates of (31) were found to be transparent in the carbonyl region.

The decrease became less as the group R became bulkier, since the in

creasing steric hindrance meant that the bicyclic form was less possible.

Since V for (23) was higher than that reported for "the ketone (32)

(1717, 1706 cm, “1)» it appeared that the opposite effect might be occurring, viz., a repulsion between the amine and carbonyl functions of (25). Unfortunately, (23) did not form a picrate. A comparison of its IR spectrum with that of (23) would have been of interest.

The IR carbonyl stretching frequencies of some R-substituted analogues of (23), when compared with that of (23), might however serve to provide evidence for amine-carbonyl interaction. This is discussed later.

Further evidence for the chair-chair conformation of the - 86 -

3-uzabicyclo (3.3.1) nonon* system (see i«fcro4uxtiQr%) W&3 provided by examination of the IR and spectra of the alcohol (33),

^9Pl5^3®5* Pr®P^®d in high yield by sodium borohydride reduction of the ketone (23). Very strong intramolecular hydrogen bonding in

(33) was apparent from its IR spectrum (concentration independent absorption at 3450— ca. 3100 cm. ^). Moreover, the N-nethyl peak at 2822 cm.“^ was at the highest frequency observed for any compound of this type.

Both these facts illustrated the endo configuration of the 7-hydroxyl group. Confirmation of this was derived by measuring (in CDCI3/D2O) 52 the half-band width of the M R multiplet at 5*80 (lH) due to the Cj carbinyl proton. The value of llc/s showed the Cy proton to be equatorial, i.e., the hydroxyl group was axial (endo). In agreement with this, the C5 and Cg protons resonated as a doublet at 7*63 (4H), with J«3c/s, the value expected for axial -equatorial and diequatorial coupling. In CDCI3, the multiplet was much broader (J-base 30c/s), due to additional coupling with the hydroxyl proton. Indeed, the 59,60 hydroxyl proton gave rise to a doublet at 2*9 (lH, J=12c/s), indicating that this proton was not undergoing rapid exchange, but was constrained by tiie nitrogen atom. The value of J pointed to a dihedral angle of ca. 180° between the carbinyl and hydroxyl protons, which agreed exactly with a model of (33) with the hydroxyl proton directed right at the amine nitrogen atom. w) •

At this stage, it was thought that the projected investigations - 87-

on further extensions of the reduction — Mannich sequence could b© combined with a search for a synthetic route to ^--substituted analogues of the ketone (23), in order that comparisons of carbonyl stretching frequencies could be made.

Accordingly, a series of amines were substituted for methylamine in the Mannich step. m-Binitrobenzene, on reduction and treatment with formaldehyde and ethylamine, furnished 41$ of a basic dark red oil, from which was obtained a low-melting crystalline solid. This material was shown to have the expected structure (34) by analysis (C^QH^MaO^), mass spectrometry (P= 241 m/e), and IR absorption at 1555 and 1353 cm.“^ (nitro). As in the case of (la), the two olefinic protons gave rise to a sharp M R singlet at 3*98, and the allylic protons to a broad singlet at 7*26* The remaining signals due to the bridge, N-ethyl, and NCH2 protons were all assigned (Table 16).

The use of benzylamine led to the formation in 12$ yield of a basic dark red oil, which could not be obtained solid. It was shown however by its IR absorption (nitro, aromatic C=C and =C-H bands) to be the N-benzyl compound (35), and was characterised by its picrate, which analysed correctly forC2iH2(>N50]_]_. The MIR spectrum of the picrate exhibited similar olefinic and allylic proton signals to those of (la) and (34), (vide supra).

When ammonia was used as the amine in the Mannich step, the -88-

basic fraction, obtained in 40$ yield, was a dark yellow oil which

partially solidified. The routine IR speotrum showed N-H absorption,

and TLC, conducted in an amine buffer solvent, indicated the presence

of two amines. By washing the oil with hot solvents, a straw-

coloured solid with m.p. 132-140° which could not be further purified

by recrystallisation, was obtained. This material had V 1562 BlclX and 1358 cm.“^ (nitro), but no N-H str. band. However, it was later

observed that other fully characterised N-H compounds of this type

(vide infra) exhibited no band in the N-H str, region. This product

was probably crude (36), since the NMR spectrum contained a singlet at

1*65 (lH, amine proton), a doublet at 2*46 (2H, J=6c/s, olefinic

protons) and a broad singlet at 6*37 (NGH2 protons). Its picrate had a wide m.p. range, which appeared to include two melting points.

Indeed, TLC analysis of the picrate confirmed the presence of two

compounds, but lack of solubility precluded the use of preparative TLC

for their separation.

Unsuccessful attempts were also made to synthesise the N- phenyl and N-acetyl compounds (37) and (38) respectively, by substitution

of aniline and acetamide for methylamine. In the former case, the small

basic fraction contained aniline, and a mixture of three other compounds.

In the latter case, the product, a mixture of two compounds, did not have

IR absorption expected for a tertiary amide.

In contrast to m-dinitrobenzene, however, methyl 3,5- - 89-

dinitrobenzoate, after sodium borohydride reduction and tffeatment with formaldehyde and ammonia, afforded a good yield of a single crystalline compound, the methyl ester (39) • This compound analysed for jNjOg> and had V ^ ol 1736 (ester 0-0), I664 (C-C), and

1276 cm. (0*0), in addition to nitro group absorption. Absorption due to N-H was however absent. Its NMR spectrum was foupd to be extremely similar to that of (2c), (both in CF^CC^H, Table 21), with a singlet at 1*64 (lH, amine proton) instead of one at 6-7i

(3H, KCHj). A broad singlet at 2*75 (2H) and a sharp one at 6*02

(3H) were due to the olefinic and methyl ester protons respectively.

Hy hydrogenation of (39) to the endo ester (40), followed by thermal cyclisation, it was hoped to synthesise the azaadan- antane lactam (4l)> formation of which would constitute added proof of the double chair conformation of the 3-uzabicyclo (3*3*1) nonane system (cf, the preparation of the alcohol (33))* Hydrogenation of an acetic acid solution of (39) over Adams* catalyst yielded a colourless viscous gum, which slov/ly darkened on standing, and was too insoluble for TLC analysis. Routine IR showed the absence of the double bond peak at 1655 cm.”l, but the presence of an unassigned band at 1635 cm."1 Ester and nitro group absorption still remained.

A nujol mull of this material was maintained for a short time at a temperature of 180-200°, which caused disappearance of IR ester carbonyl absorption. No IR band at ca, 1680 cm.“l expected for the lactam (41) -90-

was however present. Removal of nujol left an intractable brown gum.

Similar reduction - Mannich treatment of 3»5—dipitrobenzamide furnished the bicyclic amide (42). So insoluble was this compound, that the analytical sample, C9H12N4O5 , was prepared by washing the high melting (234-235°) solid with hot solvents. The unsaturated amide group was identified by IR (nujol) bands at 3458, 3410

(free N-H str.), 3190 (bonded N-H str.), 1680 (amide I CM)), 1602

(amide II C=0) and I652 cn.*”^ (C=C str.). Absorption at 3525 was tentatively assigned to the secondary amine function. It had been hoped that the desired aminoketone (45) would be available by application of the Hofmann reaction to this -amide, but due to the extreme insolubility of the amide, this conversion could not be attempted.

It thus appeared that the sequence was of use only with primary amines such as methylamine, ethylamine and benzylamine, and in certain cases, with ammonia.

The preparation in good yield of the amide (42) from 3>5- dinitrobenzamide augured well for the syntheses of the amides (43) and (44)9 transformations of which to the aminoketones (46) and (47) were planned. However, the attempted preparation of (43) gave a crude red solid, which, by virtue of its IR spectrum, certainly contained the desired amide, but which resisted all attempts at recrystallisation. The amide (44) was not obtained pure. The usual -91-

procedure afforded a yellow oil, which although, exhibiting the IR

absorption expected for (44)» consisted of a complex mixture, and could not be purified.

The failure of this approach to the aminoketones (46) and

(47) necessitated a search for an alternative route. It was envisaged

that application of the aforementioned extensions of the Mannich step

(to ammonia and other primary amines) to 3>5-dinitroanisole would render available the vinyl ethers (48)> (49)t and (50), hydrolysis of which would afford (45)> (48) and (47) respectively. The feasibility of this

sequence was firstly tested by preparing the vinyl ether (2e). This rather unstable crystalline compound, C10H15N3O5, possessed IR and NMR absorptions similar to those of other compounds of type (2),

(Tables 1 and 15respectively). Quantitative hydrolysis occurred on heating with dilute acid, affording material identical to the ketone

(23) derived from the amide (2d), This route to (23) was far superior to that previously followed. This ready hydrolysis of the vinyl ether

(2e) contrasted with the anomalous stability of its isomer (id).

With ammonia in place of methylamine, 3»5-dinitroanisole yielded a basic fraction containing two compounds, the less polar of which predominated. After separation by preparative TLC, the major product was found to be the N-methyl vinyl ether (2e), previously prepared.

This was presumably formed by methylation of the desired N-H analogue

(48) by formaldehyde. The more polar constituent, a pale yellow oil which -92-

did not solidify, was obtained in 7% yield# Comparison with, the spectral characteristics of (2e) confirmed the presence of the desired compound (48)# IR peaks were observed at 3400 (liquid film, lf*-H str.), and at 1663 (C=C), 1550, 1366, 1339 (N02) and 1236 («C-0, all lr>

CHClj) • A carbonyl impurity band at 1740 cm, ^ was also present,

A singlet at 7*93 (lH) in the MIR spectrum was attributed to the amine proton.

Yields of 1 0% and 24% of the N-ethyl (49) and N-benzyl (50) vinyl ethers respectively were realised by a similar route involving ethylamine and benzylamine. Neither of these rather unstable yellow oils was characterised analytically, but structural confirmations were carried out by means of IR and NMR spectroscopy, Their IR spectra contained all the absorptions mentioned above for (48), except for a peak at ca, 3100 cm,~^(=C-H str») instead of that at 3400 cm.~^ The

NMR absorptions of the four vinyl ethers are assembled together in

Table 23. In each case, the olefinio and allylic protons appeared as broad singlets at ca, 5*0 and 7*2 respectively, and the methoxyl protons as a singlet at ca, 6*4 (3H), Complex signals at ca, 7*4—7• 7» and a triplet at ca, 6*80 (2H), were due to the NCH2 and the bridge protons respectively.

On hydrolysis with dilute acid, the crude vinyl ether (48) furnished 70% of material with a broad carbonyl band at 1735 cm,“^

(CHCI3) in addition to secondary amine and nitro absorptions. This product, which was mainly one compound, the ketone (45)t partially solidified, but resisted purification by recrystallisation.

Hydrolysis of (49) and (50) afforded the desired aminoketones

(46) and (47) in 75 and 70% yields respectively. The former ketone,

C10H15N3O5 , exhibited V ^ 4 1732. and V ^ o 5 1727 aJad the latter, C^HiyNjO^, V CCI4 1731 and v 1727 cm.“^

Comparison (Table7) of the carbonyl stretching frequencies of th* three N-substituted aminoketones (23), (46) and (47), showed that there was a negligible decrease in that order in the case of CCI4 , but none in the case of CHCI3. This strongly suggested that the amine and carbonyl functions at the 3- ^ncl 7- positions respectively in this system exerted no interaction on each other. The M R resonances of the ketones (46) and (47) were similar to that described (vide supra) for the ketone (23), (Table24), A quartet at 7*57 (2H,J»7c/s) and a triplet at 8*97 (3H, J«7c/s) confirmed the presence in (46) of an N-ethyl group. The N-benzyl grouping of (47) caused a multiplet at 2»7 (5H, aromatic protons) and a singlet at 6*32 (2H, benzylic protons),

Variation in the aldehyde used in the Mannich step was next investigated. After reduction and treatment with methylamine and benzaldehyde, m-dinitrobenzene gave a brown oily basic fraction, which was found to be a complex mixture. When benzaldehyde was replaced by acetaldehyde, the very small basic fraction, a dark gum, consisted mainly of one compound, but could not be solidified or -94-

converted into a solid derivative.

Participation of the 1,5—dialdehyde glutaraldehyde in this reaction would theoretically be expected to result in the formation of

the azatricyclOdodecane (51)* In practice, however, use of a 25^ aqueous solution of glutaraldehyde led to an almost insignificant basic fraction that was not further examined* Since the compound (2c) had been available in much higher yield th5-dinitrobenzoate as starting material in an attempt to

synthesise the tricyclic compound (52). The basic fraction isolated in this instance was however a complex mixture.

It could thus be seen that, not unexpectedly, only formaldehyde was sufficiently reactive to participate in this Mannich reaction.

By Michael condensation with electrophiles such as acrolein, acrylonitrile and phenyl vinyl ketone, the salts of type (55)* obtainable 55 by reduction of the corresponding m-dinitrobenzenes, promised wide synthetic versatility in the preparations of caxbocyclic 1,5-dinitroO* bicyclo (5*5*1) non-2“enes, (Scheme I). Accordingly, suspensions of

the salt (55, R=H) in the solvent mixture were treated at 0° with equimolar amounts of acrolein, with or without the addition of various bases (e.g., 0H~ OEt"), After acidification, the dark oily products showed carbonyl absorption (indicating that some condensation had occurred), nitro, and hydroxyl absorption (perhaps due to cyclisation of the side chain). In the absence of ba^se, considerable amounts of m-dinitrobenzene were recovered, together with two or three other compounds. Using base and -95-

long reaction times, the products contained no m~dird.troben»ene, hut were complex mixtures, distillation of which caused decomposition#

Isolation of the salt (53* R^H) prior to addition of acrolein gave no purer products. The dry salt was also found to explode when touched.

Similar results were encountered with acrylonitrile, viz., undistillable oils with IE nitrile, nitro and carbonyl absorptions, and containing several constituents.

Using phenyl vinyl ketone and the salt (53* R*H), (previously isolated and washed with dry ethanol) under a nitrogen atmosphere at a temperature of -70°, aliquots were removed from the reaction mixture after 5*20, and 40 minutes and compared with the final product

(isolated after 1 hr.). All had identical extremely complex thin layer chromatograms. IR spectroscopy indicated that condensation had occurred, since aromatic double bond -and carbonyl, and nitro, bands were present. None of the products could however be obtained solid.

These experiments demonstrated convincingly that the salts

(53) were of no synthetic value in the Michael reaction under these conditions.

By acidification of the disodium salts (53)* Severin et al. obtained dinitrooyclohexenes, e.g., l-methyl-2,4“'dinitrocyclohex-l-ene

(54) from 2,4-dinitrotoluene. It was felt that Michael reactions, which were unsuccessful under the conditions described above, might proceed smoothly from these dinitrooyclohexenes in the presence of base. - 96 -

Repetition of the published procedure for the preparation of (54) yielded after distillation a red oil exhibiting IR absorption at

1685 (C=C) and 1560 cm*“^ (RO2). It was however an unstable mixture of three compounds, two of which predominated (perhaps double bond isomers) * On standing, V was observed to diminish max considerably.

Since the dinitrocyclohexene acid (55) w&s reported as a solid, m.p, 155°» by Severin , it was decided to attempt the prtparation by a similar procedure of "the isomeric acid (56) from the readily available 3»5~dinitrobenzoic acid. The red oily acidic fraction, shown by TLC analysis to be a single acid, decomposed on standing to material insdfcu-ble in ether. Also attempted was the synthesis of the ethyl ester (57) of (56). Ethyl 3,5-dinitrobenzoate afforded by similar means a dark red liquid, with IR ester, nitro and double bond absorption, but comprising several compounds.

Owing to the impurity and/or instability of these products, no condensations with electrophiles were attempted.

Hannich bis- condensations with formaldehyde and methyl amine at sites activated by groups other than nitro, such as carbalkoxy, carboxy, and amido, are well known (see introduction). Examples of condensations with compounds possessing carbethoxy groups are the conversions of 2-carbethoxycyclohexanone, 2,5-dicarbethoxycyclohexanone 18,20 24 63 and 2,4-dicarbethoxycyclQ£entanone to (58), (59) and (60) respectively. - 97-

It was planned to convert the readily available 2—coxbethoxy—

cyclohexanone to the compound (58) , which on LAH reduction would

afford the diol (61). Tosylation of (6l) was anticipated to provide

the monotosylate (62), oxidisable by Jones reagent to the ketotosylate 61 (63). By analogy with other bicyclic ketotosylates, e.g., (64),

base-catalysed bridge fission of (63) was expected to yield the

azacyclooctane (65)• A similar sequence from 2-carbethoxycyclo-

pentanone via the ketoester (66) to the a*acycloheptane (67) was also

planned. 18,20 The literature procedure for the preparation of (58) gave

a mixture of four compounds, from which reasonably pure (58) was

isolated by preparative TLC in 15-21$ overall yield. IR absorption

occurred at 2790 (H-methyl), 1741 (ester 00) and 1725 cm. “-^(ketone C=0). 20 The corresponding literature values were 2770, 1730 and 1718 cm.--1

respectively. The mass spectrum showed the correct molecular weight

of 225, with peaks at 224 (loss of H), 196 (loss of Et) and 180 m/e

(loss of OEt). Its M R spectrum, which has not been reported, was 24 very similar to that of (59)» (Table25).

Under similar conditions, 2-carbethoxycyclopentanone

furnished a much lower yield of basic material, which comprised four

compounds. After separation by preparative TLC, examination of the

two main components showed that neither was the desired ketoester (66)•

Exposure of the ketoester (58) to excess LAH in refluxing

THE provided 82$ of the diol (6l) as a viscous oil which slowly solidified. -98-

Re crystallisation gave a colourless solid with a wide n.p* range,

105-131°* Both C9 epimers were therefore probably present.

The m.p. could not be raised further, and satisfactory analysis figures were not obtained* Mass spectrometry however indicated the correct molecular weight of 185, and peaks at I84, 168 and 154 m/e corresponded to losses of H, OH and CH2OH respectively. At this point, it was considered that spectroscopic characterisation of the diol (6l) would be rendered more convincing by comparison with the 6l 61,62 61 available known diols (68), (69) and (70). Their V values are listed in Table §♦ A sharp peak at ca. 3630 cm.-l was assigned to the iree hydroxyl frequency, and broad, concentration independent absorption at 3540 cm. ^ to the intramolecularly hydrogen- bonded hydroxyl frequency. In the M R spectra (Table 26), the hydroxyl protons gave rise to a broad singlet in the region 6•9-7*8» the bridge carbinyl proton to signals of different types at ca. 6*3$ and the other carbinyl protons to either a singlet or a doublet at ca.

6*6. In the case of (61), a singlet at 7*89 (3H) was attributed to the N-methyl protons.

When treated with tosyl chloride in pyridine at 0°, the diols (68), (69),(70) and (6l) formed the monotosylates (71) > (72),

(73) and (62) in yields of 99» 64, ca. 73 and 10$ respectively. In “the cases of (71) and (72), preparative TLC was employed to separate the C\q and C9 epimeric pairs respectively* Reasonably pure (62), a brown oil, was obtained also by preparative TLC* The

IR spectra of the monotosylates (Table 9) exhibited peaks at ca. 3630 (free OH)* 3570(broad* concentration independent absorption,

(intranolecularly bonded OH)* 1370* 1192, 1180 (S-0), and 1600,

1500 cm.“^(aromatic C=C). Similarities were also observed in their UMR spectra (Table 27). The tosyloxymethyl group was responsible for a pair of doublets at ca, 2*0 and 2*5 (4H, each J ca. 9c/s, aromatic protons), an AB quartet in the region 6•0-6*4 (2H, J=9-70c/s, non-equivalent methylene protons) and a singlet at 7•48-7•58 (3H, aromatic methyl protons). The single bridge carbinyl proton resonated as a* singlet at ca. 6*3 *

An epimeric mixture of the monotosylate (72) has been reported as having m.p, 65-67°. In the present work, the less polar epimer was found to have m.p, 71-72°, and to analyse correctly for C^q^^/jS. The other epimer was an oil which could not be induced to solidify. Similar behaviour was exhibited by the two epimers of (71) > "the less polar halving m.p. 130-137°. The monotosylate (73) analysed correctly for C1^H28°4S#

Jones oxidation of the less polar epimer of the monotosylate

(71) afforded a mixture of two compounds. The less polar of these, obtainable in 57% yield after preparative TLC, had m.p. 82•5-83*5°t “100-

and analysed for the ketotosylate (74), I516 other constituent, present in much smaller amounts, was also a crystalline solid, and was shown by IR and NMR spectroscopy to possess the tosyloxymethyl group, but to lack a double bond or a carbonyl group* This material was not further examined, nor was the oxidation product of the more polar epimer of (71), which constituted a complex mixture containing some acidic material.

The ketotosylate (75), C18H22O4S, was formed by oxidation of either epimer of the monotosylate (72). Repeated running of the

TLC chromatogram of (75) revealed the presence of two compounds with nearly identical Rf values. These were considered to be double bond isomers, formed in the acidic oxidation conditions.

The ketotosylate (76) was available in almost quantitative yield from the monotosylate (73) • It had the expected molecular weight of 350 (mass spectrum). Major peaks in the mass spectrum were at 195 (loss of Ts), 179 (loss of OTs) and 178 (loss of HOTs) ci/e.

The ketotosylate (63) could not be obtained in a satisfactory state of purity, despite the use of preparative TLC. It was a brown gum which did not solidify.

Table 10 shows the absence of IR hydroxyl peaks, and the retention of bands due to S=0 and aromatic C=C, similar to those described above for the diol-monotosylates. The bicyclo (4*3*1) decanes (74) and (76) had V CjO qyog and 1709 cm."1 respectively, -101-

and the bicyclo (3.3.1) nonanes (63) and (75) had V c~° 1715 max and 171S cm,“* respectively. The NMR spectra (Table28) demonstrated the absence of the bridge carbinyl proton. The only other ifference compared with the spectra of the diol-mono tosylates w-as that the methylene protons of the tosyloxymethyl group, with the exception of those of (76), no longer resonated as an AB quartet in the region

6»0-6*4, but as a singlet in the region 5• 8-6*0. It was therefore concluded that in the absence of the constraint imposed by the intramolecular hydrogen bond in the diol-mono to sylates, the tosyloxymethyl group of the ketotosylates was free to rotate, resulting ir equivalence of the two methylene protons. This transformation from non-equivalence to equivalence was manifested by replacement of the

AB quartet by a singlet.

Now that the four ketotosylates were available, the action of strong base on them could be examined. However, since the ketotosylate (63) was not obtainable pure, it wss decided to investigate only the behaviour of (74)» (75)> and (76) under basic conditions.

On prolonged reflux with ethanolic ethoxide, the ketotosylate

(76) was partially converted into two other compounds. The more polar of these products had V jnax^ 3&30 (free CH), 3515 (concentration independent peak, intramolecularly bonded CH), and a doublet at 1696,

1686 cm.~l (carbonyl). These data suggested the presence of the ketol

(77), a possible mode of formation of which is outlined in Scheme J. The twin carbonyl was explicable on the basis of a hydrogen-bonded

(1686 cm. ““•*-) and a non—hydrogen—bonded (1696 cm* carbonyl group.

Treatment of the more polar product with tosyl chloride in pyridine effected partial conversion to a less polar compound with a Rf value identical to that of the .ketotosylate (76). This confirmed the formation of (77) •

The ketotosylate (75)» after subjection to a four day reflux with ethanolic ethoxide, afforded a mixture containing unchanged storting material. IR absorption at 3500 (hydroxyl) and 1710 cm.”-^

(carbonyl) however suggested the presence in the mixture of the ketol

(78).

With t-butoxide in t-butanol, (76) was converted completely into a mixture of four compounds. Since "the IR spectrum of the product was similar to that of the ketol (77)> it was probable that the most polar constituent was (77).

Similar IR absorption was exhibited by the products from v74) and (75), and it again seemed likely that one constituent of each of these mixtures was (79) and (78) respectively.

None of the products from these attempted bridge fission reactions was observed to have IR peaks due to an ester carbonyl or an exomethylene group.

The remarkable resistance to bridge fission exhibited by the compounds (74), (75) and (78) necessitated some explanation. “ 103 -

Examination of models of these ketotosylates showed that the

tosyloxymethyl group was free to rotate, a fact which was mentioned

when the NMR spectra of these compounds were discussed (vide supra)•

Hence in only one position (80) is the CH2-OTs bond in the required

trans and coplanar relationship with the bond linking the bridge

carbonyl with the carbon atom carrying the tosyloxymethyl group.

Any other position e.g., (8l),of the CHp-GTs bond would be expected to

prevent bridge fission and concomitant elimination of tosylate anion. 61,62 In ketotosylates, e.g-, (64), reported to undergo bridge-fission,

the two bonds cleaved were fixed antiparallel to each other. Table 1 Absorption frequencies in CCl^ of some 6- and 7-substituted l,5-dinitro-3-methyl-3-azabicyclo(3.3.l)non-6-enes A; 0

C\J str. absorption due NMe K asym. sym. 1 to R H 2790 3 048 (=C-II) 1547 1354,1338 795,75 2^ ; i Me 2802 1553 1343 1 fU i ! 2802 1665 1562 1568,1343 I ' 01 ON NO, 1552 A s K ' OMe 2804 1680 1561 1545 1245(«c-o) 1 1553 Me 1 C0oEt 2804 1660 1560 1734sh,1723(0=0) c 1361,134a 1554sh 1262(0-0) C0oHa 1655 1560 1356,1344 2260-2760C(MH+ 1546 ionised aminoacid6) ! 1708(0=0) j

C02Ha 1657 1555 1349 2330-2630d(NH+ \ Q , | B ionised aminoacid ) i 1703(0=0 ) A a c o n h 9 2808 1659 1551 13p2 3458(free KH),3170 j r 2 ON P .(bonded KH),I696 | n (amide I C=0),l6l8 ! SA (amide II 0=0) j 1 I Me C02Et 2800 1660 1553 1368,1341 1724(0=0 ),1272, | 1261(0-0) COpMe 2798 • 1661 1555 1568,1341 1729(0=0),1274, 1262(0-0) OMe 2800 1669 790,7401 1552 1369,1342 1235(=C-0)

a. nujol mull b. CSp solution c. broad absorption, with peaks at 2700, 2635, and 2570 cm.”^ d. broad absorption e. Nakanishi, Infra-red Absorption Spectroscopy, p. 39. f. thin film / Table 2 Absorption frequencies in nujol of some 6- and 7-suhsiitutea 1,5-dinitro-3-methyl-5-azabicyclo(3*3*1)nonanea

HO 2 s t r • R? M e absorption due to A R1 K3 asym. sym. 1 . .j H H Ha 1362 R, 2797 1345 1 H OH OH 2808 1551 1353 3467sh.arp(free Oil) JK 1348 R/ 3310broad(bonded CH; 1 o 2n n o 2 Me OH OH 2613 1549 1331 3515,3495 s k 1533 1 Me H OH 2808 1551 1354 3625 ; M e HH 0Hb 2822 1351 1342 3450-ca.3100c

a. CCl^ solution b. CHC1-. solution c. broad absorption, unaffected by dilution,i.e., intramolecular hydrogen bonding

Table 3 Absorption frequencies in nujol of some 7“substituted~3- azabicyclo(3.3*l)non-6-enes

HO 2 str. R NH C=C absorption due to R asym. sym • H 1562 1358 R COMH, 3525 1652 I56O3I1 1348 3438,3410(free amide 1551 HE),3190broad(bonded amide EH),l680(amide o2n NOz I C=0),1602(amide II .8=0) COgM € 1664 1568 1346 1736(0=0),1276,1230 1556 H (c-o) 0Mea 340Gb I663 1550 1366 1236(«C-p) 1335

a* CHCl^ solution b. liquid film Table 4 Absorption frequencies in CC1/ of some ^-substituted an. 7-substituted 1,5-dinitro-5-azabicyclo(5.5.1)non-6-enes

C=C NO? str. , , , -q R R absorption due to R str. asym. sym. H ft 1)55 155)

R H CllpPh 1jyb 5060,5050(ar.=C-H str.) l606,1498(a r *C=C str.) 748,706(ar.=C-H def.) Q.N 2 MQ* OMe Et 3097(«O-H)L353 1570 l?55(»C-0) 1540 S k OMe CH„Ph 509l(=C-Ii>554 1571 5070,5050(ar. =C-H str.) R 1670 1542 1497(ar.C=c) 12j?(=0-0) a. thin film

Table 5 Comparison of the carbonyl stretching frequency of the aminoketone (25) with those of cyclohexanone and several bicyclic ketones

ketone in CC1 in CHC1 c=o c=o cyclohexanone aminoketone ( 1727

1729

1709

M e (30) 1710

■Cc^Et (58)

a. Zbinden and Hall, J. Amer. Chera. Coe., i960, _62, 1215. b. Eglinton, Martin, and Parker, J. Chem. Soc., I965, 1245. c. Leonard, Morrow, and Rogers, J. Amer. Chem. Soc., 1957, _79> 5476. d. Reference 9- e. Reference 17* f. present study, g. Reference 20 h. O.S.Poote, Ph.D. Dissertation, Harvard University, 1961. Table 6 Absorption frequencies in CHC1 of some 3-sut>stituted.-l,5- dinitro-3-azabicyclo(3.3 •1)nonan~7-ones

N0? str. R absorption due to R asym. sym. i 0 A Me 1551 1369 2808 Et 1553 1370 02n CH?Ph 1552 1372 3035(ar.=C-H) A 1 1495(ar.C=C) R H 1553 1361 3400(N-H) 1

Table 7 Carbonyl stretching absorption frequencies of some 3-substituted-1,5-dinitro-3-azabicyclo(3.3.l)nonan-7-ones

R CC14 CHCl^ o A Me 1734 1727 1732 1727 ON n o 2 St Z CH2?h 1731 1727 H R I735a a. centre of broad band

Table 8 Hydroxyl stretching absorption frequencies in OCl^ of some bicyclic 1,3-diols

compound free OH str. bonded OH str.1* diol (6l)a 5635 3535 diol (68) 3655 3540 diol (69)° 3620 3500 diol (70) 3640 3580 a. C-H str. of NCH at 2780 cm."1 b. broad absorption, unchanged by dilution, i.e., intramolecular hydrogen bonding c. CHC1, solution 3 IdiP.lf 2 Aosorption frequencies in CGI, of some bicyclic diol 1 3- monotosylates ^ ' *

compound OH str. free OH bonded 0H° 8=0 str. C=C j monotosylate (62)a 3620° 3560,3475C,e 1360,1192,1180d 1601,1496 | i (ar.C=C) monotosylate ( (71) 3635f 336C 1370,1192,1180 3013(=C-H) 3640 3565^ 1498(ar.C=C)| monotosylate (72)g 3640 3373 1370,1192,1180 3025(=C-K) | 1602,1499 (ar.C=C) monotosylate (73) 3640 3380e 1880,1372,1192 1601,1497 1181 (ar.C=C) j - __■ .-I a. C-H str. of NCH^ at 2780 cm." 1 (CHC1 ) b. as above c. as above •d. thin film e. approximate values only, since absorption is very much broader than the other cases f. more polar epimer; other absorptions were identical to the less polar epimer g. less polar epimer

Table 10 Absorption frequencies in CC1, of some bicyclic l-keto-3- tosylates

compound C=0 str. 8=0 str. C=C ketotosylate (63)a 1713b 1370,1190,1180 l600,1493.(ar.C=C str.)

ketotosylate (74) '' 1708 1372,1190,1179 3020(=C-H str.),1600, 1496(ar.C=C str.) ! i ketotosylate (75) 1718 1374,1191,1180 3025(=C-K str.),1601, 1498(ar.C=C str.) | ketotosylate (76) 1709 1374,1191,1180 l600,1496(ar.C=C str.) | ! --- 1 a. less exact values, recorded on a thin film; also, C-H str. of NCH, -1 at 2780 cm. b. CHC1- solution; also, absorption of unknown origin at 1667 cm. 3 fables 11, 12, and 13 Mass spectra of some 6-substituted (Table ll) and 7-substituted (Table 12j 1, 5-dinitro-3- methyl-3-azabicyclo(3.3.l)non-6-enes, and of some 1,5 -dinitro-3 -azabicyclo( 3*3*1 )nonanes a (Table 13) Table 11 s 1 0 s s e I compound R mol.wt . NO NO HNO 2N0 2N0o HN0o+N02 others (30) (46) (477 (60) (927(93) H ??7 181 180 167 135 134 CH, 241 211 193 194 181 149 148 Cl 231 215 214 201 169 168 OzN NO, 233 217 216 203 171 170 Ssr OMe 257 227 211 210 197 165 164 243(CH ) He c o 2h 271 225 224 211 179 178 234(011) 226(CO H)

CO Et 299 253 252 207 206 254(OSt) Table 12 R CO?H 271 241 225 224 211 179 178 ^ 254(GH)d COpMt 299 269 2^3 252 239 207 . 206 254(OEt) ON NO, 2 COpMe 285 255 239 238 ■ 225 193 ,192 ?54(OMe) ■N- c o m ? 270 240 224 223 210 >178 177 Me Table 13 olefin (34) 241 211 195 194 181 149 148 226(CH,) 9l(possibly o f 7+ ) amine (l?) 229 199 183 182 169 137 136 diol (15) 261 215 169 168 ?44(OH) 243(H?0) a. common to all the spectra in these tables was a peak at 91 m/e (possibly C^H^+) Table 14 M R spectral assignments of some 6-substituted-l,5-dinitro- 3-methyl-3-azabicyclo(3.3.l)non-6-enes in CDC1.. 3

t y p e 0 f p r 0 t 0 n \ % R olefinic allylic bridge n c h 2 , NCEL others 1 3 Ha ’° 3*89s 7»20broad s 6*75^ 7-4-7-7b 7-57s J=llc/i3 complex ch5 4 *27111 7* 24m 6.60e ca.7*6^ 7.58s 8-36d half-band half-band 6*79 7-70s J=2c/s Q,N NO^ width 7c/s width 8c/s methyl o o i —i 3*86t 7-lOd 6*59® 7-59° 7-55s i j«4c/s J=4c/a 6*77 7* 60 M e OMe 5-13t 7*15d , 6*60e 7*30b 7,0 8s 6«45s J=4c/s J=4c/s 6*78 7*70 methoxyl broad s C0oEt 2.85t 7*03d 6-58® 7* 24s 7.59s 5-79q,8*73t C J=4c/s J=4c/s 6*75 7* 40s both J=7c/s ethyl ester

a. Table 18 compares the signals of this compound in CDCl^,. , and CF^CO^ b. partially obscured by the NCH^ resonance c. picrate showed no shift in signals d. apparent triplet e. both broad singlets f. the bridge protons of these compounds all had this J value

Table 13 M R spectral assignments of some 7-substituted-l,5-dinitro- 3-methyl-3-azabicyclo(3«5.l)non-6-enes in CDCl^

type of proton

R olefinic allylic bridge liCHp NCH. others 7*26° 7-6ls 5-75q,8-67t C02Et£ 7*01d 6-77 6c/s 6c/s J=llc/s 7-40 both J=7c,/s ethyl ester NOj, C02Me 2*80® 6-96® 6*72b 7-24° 7*6ls 6-17s 6c/s 6c/s J=10.5c/s57*59 methyl ester • r\3 CO DM® 5.00® 7*24® 6-85b 7.4-7.7 7 C\ 6‘45s 4c,/ s 4c/s J=12c/s complex methoxyl a. picrate showed no shift in signals b. apparent triplet c. both broad singlets d. half-band width value e. broad singlet f. partially obscured by NGH^ resonance Table I6 NMR spectral assignments of some 3-substituted-l,5-dinitro- 3-azabicyclo(3.5.l)non-6-enes in CDC17

olefinic allylic bridge NCH

Me 3 *89s 7*20e 6 • 75d 7-4-7 * 7a 7-57s J=llc/s complex ON 2 2 Et 3 *98^ 7*26b,e 6.73d 7* 44s, ca.7.4q,8*94t S s r J=llc/s 7*62s both J=7*5c/s CI-I0PhC 3*95s C 7-26e 6.73d 7.36s 2* 72s(aromatic J=10c/s 7*54s 6 * 28s(benzylic

Table 17 NMR spectral assignments of some 3-substituted-l,5-dinitro- 3-azabicyclo(3.3.l)non-6-enes in CF^CCuH 5 2

t y p e 0 f proton

R olefinic n c h 2 NR

H 2-46d 6-57® 1.65s J=6c/s z O

ON CM Me 3-75d 6*68s 2 6-79® J=6c/s N \ r 1 Et 3 *70d 6- 78e 6•28q,8 *43t R J=6c/s both J=7c/s

a. partially obscured by the NCH^ resonance b. overlapped by the NEt quartet c. spectrum of picrate (other signals at 0*S4s (2H, picrate anion H), 3*2m (1H, NH+)) f. speculative assignment

d. apparent triplet g. Cs proton signal (j=2»3c/s)v . , 1 ^ 6 Msuperimposed on a e. broad singlet h. proton signal '(j=3c/s) )broad multiplet

Table 18 NMR spectral assignments of 1 ,5-dinitro-3-methyl-3-azabicyclo (3.3.l)non-6-ene (la) in CNCl^, CgD^, and CF^CO^I solvent olefinic allylic bridge NCH NCH., 2_ CDC1. 3-8Ss 7-2(T 6•75 7-4-7-7C 7-57s J=llc/s complex ON 4-43d? 7.90e »f 8.28s 2 h 4-67t . 2a»^ N s r CFx C09H 3 • 75 d 6 *68s Me ^ J=6c/s 79e Table 19 NMR shifts'^ (in ppm relative to the parent amine) of protons ______attached to carbon atoms a and [3^ to nitrogen in amine picratesC A( u-CH2) A(fiCH) triethylamine 0-75 0-53 tri-n-butylamine 0-78 0* 21 ) diethylamine 0*54 . 0.24 di-n-butylamine 0*52 0-14d piperidine 0*48 0.23 a. solvent CDC1, 3 b. shifts for protons on£c were negligible c. picrates of primary amines, e.g., cyclohexylarnine, n-propylamine, and n-butylamine, and of some secondary amines, e.g., morpholine, were too insoluble in CDC1, d. approximate, since signal due to [oCH was a broad multiplet

Table 20 Comparison of the NMR shifts (in ppm relative to triethylamine) of triethylamine picrate and hydrochloride in CDCl^ and D^O

A( aCH2) A(pCH) o « o 1—1

CDC1, fO\ • 3 D2° D2° picrate 0-75 0*65 0*33 0-28

hydrochloride 0*64 0-67 0-38 0.28 -icfb 1 e 21 NMR spectral assignments of some 1 j5-dinitro-3--azabicyclo (3.3.l)nonenes in CF,C0oH 5 2 1 type 0 f p r 0 t0 n olefinic NCH? NR others

diol (15) 6.71s vinyl ether (2e) 4* 88s 6.84b 6‘70s 6*29s (methoxyl) acid (le) 2*31t 6 *5-6 *9 6.66s J=4c/s complex acid (2a) 2-55b 6-52° 6.67s A c /s '* 6-79 ethyl ester (2b) 2-65b 6-50° 6.67s 5 *51q,8*55t 6c/sa 6-81 both J=7c/s(ethyl) methyl ester (2c) 2-73b 6.58° 6*72s 6.03s (methyl) 4o/ s 6*85 methyl ester (39) 2-75b 6*58b 1 «64s 6.02s (methyl) 5 c/s 6 - 7-6-9 complex amide (2d) 2*84b 6*46° ■ 6.65 s ca.2*4)broad absorp 5o/sa 6* 78 tion .(amide H) amide (42) 2*92b 6*50° 1 «60d 2.42,broad absorp 6c/sa 6*78 4=5 *5c/s tion (amide H)

a. half-band width value b. broad singlet c. both broad singlets'

Table 22 NMR spectral assignments of some 1 ,5-dinitro-3-methyl-3- azabicyclo(3.3.l)nonanes in CDCl^

type of proton bridge nch9 nch^ others amine (12) 6 *47 »6 •6'p1 7.43s 7*66s . diol (19) 6*601 7.56a 7*64s 5*45mb(C_ carbinyl H) 7-73 7»27s ,8«4ls(hydroxyl H) broad s 8.58s (methyl H)

alcohol (26) 6*65q 7-5-7-9a 7-65s 5*45m<^(87 carbinyl H) J=llc/s complex 8«95b(J^dc /s ,methyl H)

alcohol (33) 6-52h 7•42a 7-49s 2*9dC ’e( J=12c/s,hydroxyl II) 6*62 5*80m (C^ carbinyl H)

7»65d®(J=3c/s,C^ and CQ H)

a-i. see next page cl. partially obscured by the NCBf resonance 5 b. centre of very broad multiple* U base ca.25c/s); half-band width value indicated equatorial (exo) hydroxyl (see discussion) c. signals disappeared on adding D O d. centre of very broad multiplet (Jb ca.jOc/s); half-band width value indicated equatorial (exo) hydroxyl (see discussion) e. doublet caused oy the hydroxyl H not being exchanged, i.e., con strained by the N atom (see discussion) f. half-band width value of 11c/s meant that hydroxyl group was axial (endo) (see discussion)

g. J value pointed to an equatorial V carbinyl H (see discussion) h..apparent pair of very slightly split doublets, but probably an A3 quartet with J=12c/s i. apparent triplet with J=12c/s

Table 23 M R spectral assignments of some 3-substituted-l,5**dinitro- 7-methoxy-q-azabicyclo(3.3.l)non-6-enes in CDCl^

t y p e o f p r 0 t 0 n R olefinic allylic bridge nch9 NR methoxyl 6.87s 7*93s . s H 4#93d 7*17b 0*71 6 36 O M e 3 c/s 4c/ s 6 * 95s A . s Me 9 *00d 7#24d 6*85° 7.4-7.7 7 62 6 *43s ON N0o 4c/s 4c/s complex 2 2 Et 4*98^ 7.25a> 6*77° 7*4-7* 7b7 * 37a 6 »40s i 4c./s complex 8*94t ft J=7c/s CH0Ph 6 »82c 6 2*74s 6*38s w 4 ‘96& d7*25ad 7-3-7* ca.6c,/s 4 c/s complex (ar.H) 6.31s

-• ■ (benzyl ic H) a. broad singlet b. partially obscured by the quartet at 7*37t c. apparent triplet, J=ll-12c/s / d. half-band width value ?A spectral assignments of some 3-substituted-l,3-dinitro 3-azabicyclo(3 *3 .1 )nonan-7-ones in CDG1?

and C0 bridge NCI! 6 8

Ke 7* 04s 6*85a 7-- 21s,7*40s 7-59s Et 7*09s 6 *82a 7*'10s,7*37s 7*36q,8.97t ClN both J=7c/s 2 CH2?h 7* 10s 6-83a 7*■24s ,7*43s 2 *7 m(J Oc/s, bas 3 aromatic H) 6 • j2s(benzylic Ji)

/ CH^CO resonance

Table 23 Comparison of the UMR spectra of 1-carbethoxy-3-methyl-3- azabicyclo(3*3*1)nonan-9-one and its 1 ,^-dicarbethoxy analogue in CUCl 3

t y p e

ethoxyl CH ethoxyl CH

5 • 77q co2et J=7c/

a 6*89s 7*65s E t O a C COudt

a. reference 24

; • Table 26 NMR s of some bicyclic 1,3-diols in CDC1

type of proton

hydroxyl carbinyl ring methylene others

a \ diol (61) 6*93“broad s 6»34m(half- 7*1-8-8 complex 7-89s(NCH..) band width ' llc/s,CH) 6*57d(J^6c/s) diol (68) 6 *90abroad s 6 -24bbroad sc 7*7-8*8 complex 4 *56m(olefinicH) 6-53d(J=3c/s.) 8 *97s(methyl) a 6 * 3 9 s diol (69) 7* 28m h ) 7-9-8*8 complex 4 -3m f 4-8 m(olef 6*55s inic IT), 8 (methyl) a diol (70) 7.8 broad s 6-02dU-e-5o/s) 7.6-C.9 complex 8 .96s(meth,,1;

6 * 7 2 3 d a • signal disappeo.red in _0 2 b. sharpened on 1) 0 exchange

c. C^q II signal d. CH2OH signal

Table P J NMR spectral assignments of some bicyclic diol 1,3-monotosylates in Ci)Cl? 5

■t y p e o f p r o t o n

hydroxyl carbinyl CHpOTs aromatic aromatic others methyl monotosylate (6?) 6*i6s 6-23 q 2-18da 7-56s 6-9-7-5(I-iCH ) J=9c/s 2*6ld 7-S9s(kch y J=8*5c/s monotosylate (71)° 8-10sd 6-20s 9 *9Bq 1*90da 7.48s 4*40broad s J=10c/s ?-40d (olefinic) J=9c/s 8 -93s(methyl) monotosylate (7?)° ca.6 *3^ 6*0 j q l-91da 7*48s 4•15m,4'65m J=10c/s 2»40d (olefinic) J=9c/s 8.97a(methyl) monotosylate (73) b . 2«18da 7-58s j=9c/s 2.63d J=8 *5c/s i a. due to two aromatic H adjacent to 0; other doublet caused by the other two aromatic H b. obscured by the CH^OTs quartet c. less polar epimer d. signal disappeared in D^O

Table 28 ' NMR spectral assignments of some bicyclic l-keto-3-tosylates in CDC1 2 ______t y p e o f p r 0 t 0 n CHpOTs aromatic ar.methyl others ketotosylate (63) 6 -Olbroad s 2.23da 7*57s 7.91s (NCH3) 2*68d , J=8*5c/s' ketotosylate (74) 5*83s 1.90da 7.48s 4.0-4*3ni(olefinic H) 2*39d 7-65d(J=3*5c/s,G0H) J=9c/s 8.92s(methyl) ketotosylate (75) 5-853 l*90da 7*48s 3 •95m?4,55^(oiei’in- 2-39d ic El) ,8 *9 2s(methyl) J=9c/s ketotosylate (76) 6*13q 2«23da 7.57s J=9c/s 2 *63d J=3 *5c/s a. as above o 2n NO' 0 2 N NO N Sr c h 3 c h 3 1 2 a R = H a r = c o 2h b R = CHo b R = C 0 2Et c R = Cl c R = C0 2M e d R = OC H 3 d r=conh2 e R“ C02H e R = OC H 3 f R = C02Et

CHo I J

h 2o c AK OoN NO 0 2 N | ^ / NO o 2n NO * r V 1 CH 3 c h 3 CH

3 4

I SC H EM E A

h 2oc 0 NO 0. N V' A) CH

° 2N^ t j N02 CH20/MeNH2 ■N > 5 CH

/ / SC H EM E B e © 0 Na' © 9 © o 2n NO OoNVVNO N a BH ^

NO n o 9 © 2

02N(CH2)2CH(CH2)2N 0 2

. n o 2

OoN NO. r^2 o2n ^ n o 2 c h 3

9 10

o 2n NO.

*N CH CH.

11 12 SCHEME C

AN r ^ C H O o2n NO' o2n 02n NO. T o) T fch3 ch3 CH 1a 13

O 2 N NO N 'L c h 3

1A 15 16 SCHEME D OH OH OHC NO ■> HO-N=0 NO. CH.

OH OHC OH + NO.

( 1 7 % ) 0 'N ch3 SCHEME E

H 2°'\0HCX *r- f +0,N r'} (15) * O^Jy i NO. SN/^2 "UO; 0<— N II N c0 CH CH. 17 SCHEME F

OHC 16 o2nY CH

OH CHO h o s ^ K CK" NO M e - " NO OoN O o N no2 02n ^s m / 1 'N' I}]- y CH ch3 C H 3

18 19 20 0 OH

R R o 2n NO. 02N NO ‘v m /' tf y N c h 3 c h 3 C H 3 21 R = NH2 23 24 22 R=OAc OH c h 3-.

o2Ni o 2n NO. T ^ N c h 3 CH 25 26

CONH

or ‘1ST CH 3 28 29 30

SC H E M E G 0e

f\i® ¥ / v s R R R 31

0 9H SCHEM E H

o 2n NO. T C 6N ^ C8 CH 3 H 32 33 C 0 2M e C02M e OoN 'NO NO y o 2n \ ^ J n 0 2 l R ■tr N 34 R = Et H H 35 R = C H 2Ph 39 40 36 R = H ' 37 R = Ph C O N Ho 38 R = C O C H

Q,N NO.

• y R 42 R= H 43 R= Et 44 R= CH 2Ph

R OCH CH 3V ^ H 0 2Cy ^ 2 Na 02Nk^HN02 02N fio. 2 1'* N 0 2

54 55

SCHEME

R n ^ ^ n o2n n o2 o2n f n o7 o 9n •NOo hf5 'OH c^h2= = ch^-c h =b

c o 2r 0 N ^ C 02Et o 2n f ^ n o 2 Y c h 3 CHo 56 R = H 58 59 5 7 R = Et

OH OH O H 4 £ •OTs

■t? "N c h 3 CH; CH 60 61 62 E = C 0 2Et Rtf

H i ' c h 3 65

OH •OH

0 OTs OH

75 76

SCHEME J

0 ■Ar > 77 ■OH

I ^ N ^ O T s c h 2

79 80

0 NM\ CH

OTs 81 -104-

EXP ERIMENTAL

general literature procedure for reaction of dinitrobenzene.3 with sodium borohydride. formaldehyde. and amine^.

Th dinitrobenzene compound in a suitable solvent

mixture was treated portionwise with solid sodium borohydride with

stirring and cooling (temperature best kept in the range

5-18°). The end of the reaction (15-30 min* for 4 gm. of

compound) was indicated by a change in the initial dark blue

colour to yellow or li^tit brown, (The salt-like intermediate

can be precipitated by isopropanol and redissolved in water).

The mixture was diluted with water, cooled, and a cooled mixture

of amine, water, and formaldehyde was added. Immediate

acidification with acetic acid, extraction with methylene chloride,

washing of organic extracts with water, drying (over CaG^), and

removal of solvents left a residue purified by re crystallisation

or occasionally by chromatography on neutral alumina with methylene

chloride.

Amendments were made to the above procedure, e.g., organic

extracts were washed always with brine, and dried always with

MgSCty. The occasions when the reaction was stirred for various

periods before and after acidification, or when chloroform was used for extraction, are mentioned individually#

The formaldehyde and methylamine used were in the form of 4C$ and 25$ aqueous solutions respectively. -105-

35 lj..3-.teitro-3-methyl-3-azabicyclo (5.3 .l)non-6-ene ft.a)

£rom m-dinitrobenzene (7 gm.) was obtained a red oil

(after overnight stirring of the reaction mixture), which slowly

crystallised. Recrystallisation from methanol, then ethanol,

gave 0-^ as a pale yellow crystalline solid (4*34 gnu; Lit. 79$). Colourless material, m.p. ’J60 (Lit. 75°) was obtained by many recrystallisations from ethanol (incurring heavy loss). Light petroleum 60-800 was later found to be a more efficient recrystallising solvent. IR, table 1 ; MR, tablel4» mass spectrum, table H,

35 1.5—dinitro~5.6-dime thyl-3-azabicyclO (3.3.1) non-6-ene Q#)

2.4-dinitrotoluene (4 gm.) formed (after overnight

stirring of the reaction mixture) a red oil, which became a

yellow solid after rubbing with a spatula. Recrystallisation

from isopropanol (Lit.) was effective only initially, but re

crystallisation from ethanol gave (Lb) as a pale yellow solid

(2-04 gm.; 36$; Lit, 58$), m.p. 97° (Lit. 98°).

IR, table 1; MR, tablel4 ; mass spectrum, table XL, 35 1.5-dinitro-3-me thyl-6-chloro-3-azabicyclo (3.3*l)non-6-ene Cks)

2.4-dinitrochlorobenzene (4 gm.) afforded (on stirring

the reaction mixture for 3 hr.) a red oil, which crystallised

rapidly on cooling. Yield of orange crystals (2X isopropanol)

2*97 gm. (58$; Lit. 44/o, after chromatography on alumina with methylene chloride). Further recrystallisations (IX isopropanol,

3X ethanol) gave (3c) as colourless crystals, m.p.113-114° (Lit. 114°)* -106-

IR, table 1; NMR, table!4; mass spectrum, tableH.

j^-dinitro-3-methyl-6-methoxy-3-azabicyclo (3*5*l)non—6—ena (Id)

2,4-dinitroanisole (2 gm.; prepared from 2,4-dmitxo- 3 8 chlorobenzene ) dissolved in THF (10 ml.), formamide (20 ml.) and methanol (10 ml.), treated under conditions described in the preparation of (££), formed a red oil (2*33 gm*), rapidly solidifying on standing. This was extracted with hot ethanol, cooled, and filtered to remove precipitated black tar. The filtrate was cooled (ice-salt bath) yielding successive crops of (Id) as an off-white crystalline solid (765 mg.; 30$). Further recrystal - isation from ethanol (large wastage) gave colourless needles, m.p.

140*5 - 141*5°.

Found: 0,46*54 ? H, 5*54 » N, 16*04* C10H15R3O5 requires

C, 46*69; H, 5*88; N, 16*33*

IR, table 1; NMR, table 1 /[: mass spectrum, tableIL.

Hydrolysis of (Li)

The vinyl ether (l )» (5 mg.), was heated for 30 min. on the steam bath with (a) dilute hydrochloric acid (6N; 1 ml.) and water (l ml.), and (b) concentrated hydrochloric acid (2 ml.), the solution turning pale yellow, then basified in both cases with saturated sodium carbonate solution. Only starting material

(IR, TLC) was recovered from (a). No product was obtained from

(b), indicating that water-soluble material had been formed. -107-

1^3-dinitro-5-me thyl-6-carboxy-5-azabicyclo (5.3.1 )non-6 -ene Js l

Application of the procedure used in the preparation of (^), to 2,4-dinitrobenzed.c acid (4 gm.; prepared from 2,4- dinitrotoluene ^ furnished (Lq) as a pale yellow solid (560 mg.;

11$), m.p. 224-226°, after recrystallisation from ethanol

(considerable wastage). The aminoacid was only slightly soluble in water, but soluble in both dilute acid and base.

Found: G, 44*29; H, 4*54; N, 15*45* CIQK13N3O6 requires

C, 44*28; H, 4 ’83; N, 15*49*

IE, table 1; MR, table21; mass spectrum, table 11«

1.5-dinitro~5-methyl-6-carbethoxy-3-azabicyclo (3.5«l)non-6-ene (if)

Under the conditions described for the preparation of ’

(2b), ethyl 2,4-dinitrobenzoate ( 2 gm,) afforded a red oil

(2*03 gm.) which did not solidify, had no R-methyl absorption in the IR, and appeared to consist of two main components (TLC).

The jrroduct was dissolved in chloroform, and washed with saturated sodium.bicarbonate solution, and dilute hydrochloric acid. The acidic fraction was a complex mixture (TLC), as were also the neutrals (constituting most of the material), whereas the basic fraction (which slowly solidified) was a single compound (TLC).

Yield 101 mg. (4$), m.p. 133-154*5°, after recrystallisation from ethanol.

Found: C, 47*77? H, 5*44? N, 13*99* QL2H3JN3O6 requires

C, 48*16; H, 5*73; N, 14*04. -108-

IR, table 1 ; NMR, table 14.; mass spectrum, tablell. l_i5-dini tro-5-metfayl-7-carboxy-5~azabicyclo(3.5.1) non-6-ene (2a)

To a stirred solution of 3»5-dinitrobenzoic acid (4 gm*) and sodium hydroxide (09 gm.) in water (40 ml.), was added a solution of sodium borohydride (0*2 gnu) in water (2 ml.). After the initially violent reaction (dark red colour produced) had subsided, a solution of sodium borohydride (1*8 gm.) in water

(8 ml.) was added, and the mixture stirred for 30 min. Dilution with ice-water (50 ml.) was followed by addition of a mixture of

15 ml. each of formaldehyde, methylamine, and water. Stirring /as continued for 30 min., the mixture acidified with acetic acid (15 ml.), and stirred for 1 hr* The usual extraction procedure afforded

as a pale yellow crystalline solid (560 mg*; 11$). The yield was not increased by a 3 day constant ethyl acetate extraction*

Recrystallisation from ethanol gave colourless crystalline material, m.p, 185-186*5°*

Found: C, 44*48 ; H, 5*07 ; N, 15*57* C10H13N3O6 requires

C, 44*20; H, 4*83; N, 15*49*

IE, table 1 ; NMR, tableSL; mass spectrum, table 12,

1.5-dinitro-3-me thyl-7-carbethoxy-3-azabicyclo( 3 * 3 * l)__non-6-ene

Ethyl 3,5-dinitrobenzoate (2 gm.) dissolved in THF (15 ml.) and ethanol (15 ml.), was reduced by sodium borohydride (1*5 gm*)*

Then were added ice—water (50 ml.), the usual mixture (7*5 ml. of -109-

each constituent), and acetic acid (7*5 ml.)* After being stirred

for 17 hr,, extraction gave (2b) as pale yellow crystalline material (2*45 gm,; 98$)* Colourless feathery crystals, m.p, 101-102°, were obtained by recrystallisation from ethanol.

Found: C, 48*15? H, 5*40; H, 13*84, C12H17N5O5 requires

C,48.16; H, 5*73? N, 14*04*

IR, table 1; NMR, table 15? mass spectrum, table 12,

Esterification of (2sj) to (2b)

The aminoacid (2h), (50 mg.), was heated for 18 hr, with ethanol (3 ml.) and dilute sulphuric acid (6 ml.), cooled, basified with aqueous sodium carbonate, and extracted with

chloroform. The organic extracts were washed, dried, and evaporated, leaving a yellow oil which contained mainly (2s), to gether with a more polar component (TLC). Separation by preparative TLC (eluant: light petroleum/20$ EtOAb) yielded

the less polar component as a yellow oil solidifying on standing

(6mg.; 11$), identical (IR, TLC) to @b) obtained above.

1,5-dinitro-3-meth.yl-7-carbome thoxy-3-azabic.yd o (3*3* 1 )non-6-ene (2c)

From methyl 3»5-dinitrobenzoate (2 gm.), following the procedure used in the preparation of (23), (2s) was obtained as a red oil which crystallised on cooling (2*47 gm.; 98$), m.p.

122-124° (colourless plates from ethanol). -110-

Founds 0,46*28; H, 5*40; N, 14*71- O u P i ^ o e retires

0,46*32; H, 5*30; N, 14*73-

IR, table 1; NMR, tablel5 $ mass spectrum, table]2-

1,5-dinitro-5«7-dimethyl-5*7-diazabic.vclo (5-5*l)nonane ( 5 )

2,4-dinitrophenol (10 gnu) in THF (25 ml*), formamide

(50 ml*), and ethanol (62*5 ml*) was reduced by sodium boro

hydride (7*5 gm*)* Ice-water (250 ml*), followed by a mixture

of 37*5 ml. each of formaldehyde, methylamine and water, and

finally acetic acid (37-5 ml.) were added. After being stirred

a black semi-solid tar on standing. Several purifications

(as for (lcj)) yielded a pale yellow crystalline solid (164 mg*;

1*2$), which became colourless needles, m.p 117-118°, on

further recrystallisation from ethanol.

Found: C, 44*34> H, 6*59> 20*95- CyH^IfyO^ requires

C, 44*26; H, 6*60; H, 22-94-

Mass spectrum (P*244 m/e) contained peaks at 243 (P-l), 198

(p-46, loss of N02) and. 151 (i5“93» loss of W 2 and HN02) m/e*

MR(CDCQ.j): 7*13 0. (J=llc/s, bridge H) and 7*34® (10H in all, NGH2) 5

7.64s (ffl, HCHj). IS (CCI4): 2812, 2798, 1551 and 1353 cm.-1

1.5.9-trinitrocyclohexane (9 /

A solution of trinitrobenzene (2 gm.) in THF (10 ml.)

was added slowly dropwise to a cooled solution of sodium borohydride -111-

(2 gm.) in water (25 ml.) and methanol (25 ml,). The dark-red solubion was tirred for 30 min., acidified with 10% aqueous tartaric acid (90 ml.) and filtered. Extraction gave a dark-red oil, which decomposed rapidly on standing to form material insoluble in chloroform. The product was extracted with chloroform to yield a more mobile red oil which did not solidify (560 mg.; 28%; Lit. recrystallised from isopropanol, m.p, 125°, 40%). It consisted mainly of one compound (TLC), IR (GHCI3) 1563 cm.~^ (R02 asym. str.).

Attempted preparation of 1.5«7-trinitro-3-methyl-3-a2abicyclo(3.5.l) nonane (lO)

Trinitrobenzene (2 gm.), reduced as in the preparation of

( 9 ) 1 yielded,under conditions described in the preparation of (2$, a red oil (1*56 gm.). Acid extraction of the product gave a basic brown oil, which appeared to be mainly one very crude compound

(TLC), exhibiting hydroxyl, nitro, and R-methyl absorption in the

IR. It could not be obtained solid. Treatment with an ethanolic solution of picric acid gave a yellow precipitate, which when filtered, quickly became a brown gum.

Attempted preparation of 1,5—dinitro-5-methyl-3-azabicycle (3*2.2.) non-6-ene m ). 4 nQ p-Dinitrobenzene (390 mg.j prepared from p-nitroaniline ) was dissolved in ethanol (3 ml.) and THE (9 ml.), and treated with sodium borohydride (0*4 gm,)« signs of reaction (cf, effervesence -112-

observed for m-dinitro compounds) were noticed, and the initial pale yellow colour became orange-pink and cloudy (NaBH^ in suspension).

The usual procedure (see preparation of (3d)) gave after two hours’ stirring a clear red solution, from which basic material was extracted as a pale brown oil (75 mgi)* The product was a mixture of three compounds (TLC), and the main constituent (separated by preparative

TLC), with an Rf value similar to that of (la), did not have IR or M R absorption expected for (ll).

1,5-dinitro-3-methyl-3-a2abicyclo (3.3*1) nonane (12)

The olefin (3§l? (227 mg.; ImM), in ethyl acetate (15 ml.) was hydrogenated for 4 hr. with 10% palladium on charcoal (100 mg,).

Filtration, and evaporation of solvent gave a viscous pale yellow oil

(188 mg.). TLC analysis showed the presence of unchanged 0a), a less polar compound, and a more polar one. Separation by preparative TLC

(thickness 1 mm., eluent: light petroleum/20% EtOAc) yielded:- i) most polar material (92 mg.), a mixture of possibly three

components (TLC), with strong N-H absorption in the IR. ii) starting olefin (6 mg.) iii) least polar material, a pale yellow oil which solidified

(85 mg.; 37%). Recrystallisation from li^ht petroleum 60-80°

gave (l2) sis colourless crystalline material, m.p. 66-69°*

IR, table 2; MIR, table22; mass spectrum, tablel3* -113-

Selenii.im dioxide oxidation of olefin (Lg)

The olefin (La), (6*84 gm.; OQ3M), in water (500 ml.) was refluxed for 18 hr. with selenium dioxide (6*72 gm.; 0*06M).

The red selenium formed was filtered, and the filtrate (acid to indicator paper) extracted with chloroform (6 X)• The organic extract was washed, dried, and evaporated, leaving a viscous dark-red oil

(1*15 gm.). The yield was not improved by either a constant ethyl acetate extraction of the aqueous layer, or a chloroform extraction of the aqueous layer after basification with saturated sodium bi carbonate solution or saturated sodium hydroxide solution. The product consisted mainly of one compound, and three less polar ones

(TLC). Occasionally, the oil could be purified by crystallisation from carbon tetrachloride, but usually purification was accomplished by preparative TLC (thickness 0*8 mm., eluants light petroieum/60%

EtOAc). The main compound, an orange oil (530 mg.), solidified, m.p. 64-680, after distillation, becoming colourless needles, m.p.

68-69°, on re crystallisation from carbon tetrachloride.

Found: C, 50*51; H, 6*20; N, 12*55* 09314^204 (214)requires

C,50*46; H, 6*59; N, 13*08, Mass spectrum (P apparently »

242 m/e, but no losses of, e.g., 30 (NO), or 46 (N02) were observed).

Also, two spectra of later products, although identical up to 170 m/e, continued to different m/e values. IR (CHCI3) 1702, 1551 cm. ^

NMR (CDCl^) 0*37s (1H, aldehyde H); 3*62d (2H, J=6*5c/s, olefinic H -114-

on cis C=C); 6*10 q (2H, 12 c/s, bridge H); 6-696 <4h ) and PXQTT 7*13 broad s (4R) could not be assigned, UV A ^ ax 227 np (e3600).

Although soluble in dilute hydrochloric acid, no picrate could be prepared, Fehling’s and silver mirror tests were both positive,

1.5-dinitro-5-methyl-3-azabicyclo (3,3,1) nonan-6,7-diol (13 )

The aminoolefin (La), (885 mg*; 3*9 mM)# in a solution 50 of osmium tetroxide (l gm.j 3*94 mM) in dry pyridine (15 ml.) was

left for two days at room temperature, A solution of sodium bi

sulphite (1*8 @a.) in water (30 ml.) and pyridine (20 ml.) was then

added. The mixture was stirred for one hour, extracted with chloroform

(3 X)* and the organic extracts washed, dried, and evaporated, leaving

(k) as a yellow crystalline solid (944 mg.j 93%)* Several re

crystallisations from aqueous ethanol gave a colourless crystalline

solid, m.p, ca. 128-144°* The narrowest m.p. range observed was

150-153° (a sample recrystallised 3 X aqueous EtQH).

Pounds C, 41*40; H, 5*49; N, 15*90, C ^ ^ O g requires

C, 41*38; H, 5*79; N, 16*09*

IR, table 2; MR, table21; mass spectrum, table13*

1.5-dinitro-3-methyl-3-azabicvclo (3*2.1)octan-6-ol (16)_

a) from diol (15)

To the diol (15), (522 mg.; 2mM), dissolved in methanol

(50 ml.) was added sodium metaperiodate (530 mg.). After being -115-

stirred, under nitrogen for 18 hr,, brine was added, and tb© mixture extracted wit chloroform. The •rganic extract was washed, dried, and evaporated under reduced pressure without heati ag (the product was unstable to heat), leaving a brown oil which slowly solidified.

Its IR spectrum showed hydroxyl, N-methyl, nitro and carbonyl

(1720 cm.~^) absorption. It contained mainly one compound, less polar than the starting diol, and an even less polar component

(TLC). (A product with no carbonyl absorption in the IR, and con sisting only of the more polar compound, was obtained by repeating the experiment with a stirring time of 60 hr.). The product was purified by dissolving in a little cold benzene, cooling in an ice-salt bath, adding a large volume of carbon tetrachloride, followed by light petroleum 40-60°, and filtering. Two such recrystallisations gave

(16) as a pale brown crystalline solid (219 mg. j 47%)» ®,p, 96-99°*

Alternatively, preparative TLC (thickness 1 mm,, eluant: light petroleum/30% EtOAc) yielded identical material from the more polar band.

The alcohol decomposed slowly on standing by elimination of nitrous acid (odour detectable), and a satisfactorily pure analytical sample could not be obtained. Hass spectrum (P=231 m/e) showed peaks at 230 (P-l), 214 (P—17» loss of OH), 185(P-46, loss of NO2), 184

(P-47, loss of HN02), 139 (P-92, loss of 2N02), and 138 (P-93, loss of

NO2 and HNO2) m/e. -116-

IR (CHCI3) 3595 Sharp (free OH), 2805, 1549, 1371 cm.'1 UMR (CBClj)

5*5 m (1H, half-band width 12 c/s, Cg-H); 6*73t (2H,J=10c/s, bridge H); 7*42, ca. 7*6 (NCH2), and 7*59S (lOH in all, NCH3); hydroxyl and Cy-H not assignable* b) from diol (L9)

A solution of diol 0-9), (138 mg.; 0*5 mM), in methanol

(10 ml.) was stirred under nitrogen with sodium metaperiodate (130 mg.) for 15 hr. Extraction as in (a) furnished a yellow oil (120 mg.) containing some starting diol (present even in a reaction mixture stirred for 2 days), and a less polar compound (TLC), The IR spectrum again showed carbonyl absorption. Preparative TLC (as for (a)) afforded material identical to that obtained from diol (15), (IR,TLC).

Attempted tosylation of (16)

A mixture containing (16), (47 mg.; 0*2 mM), dry pyridine

(1*5 ml.) and tosyl chloride (40 mg.; 0*2 mM) was kept at 0° for

16 days, and then poured onto ice and extracted with chloroform (3 X).

The chloroform extract was washed, dried, and evaporated (pyridine removed by azeotropic distillation with dry benzene) a The product consisted almost entirely of unchanged (l6)j (IP as for (16 )» with weak tosylate peadcs at 1600 cm.”^ and in the fingerprint region;

TLC showed only (l6)).

Attempted Jones oxidation of (16 )

A solution of (16)* (23 mg.; 0*1 mM), in acetone (5 ml.) -117-

was ‘treated with excess Jones reagent* Chloroform extraction yielded an oil which solidified (18 mg,). It contained only (l6 )r

(TLC), with a trace of ketone (IR, 1710 cm,“^ v.w.; remainder of spectrum identical to that of (16)).

115-dinitro-3,6-dimethyl-3-azabicyclo (3.3.l)nonan-6,7-diol (19)

Hydroxylation of the aminoolefin (itj), (470 mg.; 1*95 mM) with osmium tetroxide (0*5 gm.; 1*97 mM), in the same way as in

the preparation of 05), furnished (19) as a viscous yellow oil

(535 mg.; 100:/o) f which slowly solidified. A colourless crystalline

solid, m.p, 128-132°, was obtained by recrystallisation from carbon

tetrachloride-chloroform.

IR, table 2 ; MMR, table22* li5-diamino-3-methyl-3-azabicyclo (3*3«l) non-6-ene Ml A solution of the aminoolefin (Lg), (1*135 gm.; 5mM), in

dry THF (60 ml.) was treated carefully with a suspension of lithium

aluminium hydride (2*5 gm,) in dry THF (20 ml,), and the mixture refluxed for 4 hr* (overnight reflux gave a lower yield, whereas reduction at room temperature was incomplete). To the cooled reaction mixture was added moist ethanol, and the filtrate from suction filtration was evaporated and extracted with chloroform* The lithium salts were washed with hot chloroform, and the combined organic extracts washed,

dried, and evaporated, yielding (2l) as a volatile yellow oil -118-

(526 mg.; 63%). Routine IR (Liquid film) 3350 broad (HH), 3050

(=G-Ii), 2800 (NMe), 1610 cm*"”'*' broad (RE def.) and no NO2 str.

Distillation afforded, with 75% loss, a pale yellow oil, b.p. ca.100-

180°/ 0*3 which darkened on standing to form material insoluble in chloroform. The M R spectrum was not easily assignable, but contain-- ed 7*69s, 7*79s (ca. 4H, possibly amine H). TLC analysis showed extensive tailing, even in amine buffer solvents. Its picrate was dark yellow, m.p. ca, 130-1400, and too insoluble to be recrystallised.

Attempts to isolate a solid diacetyl derivative were also unsuccessful.

Attempted bjs-dearriination of triamine (21) a) The triamine (21), (140 mg.; 0*84 mM), dissolved in acetic acid (l ml.) was treated portionwise over 1 hr. at 0° with sodium nitrite (280 mg.). Stirring was continued at room temperature overnight, Sodium nitrite ('JO mg.) was then added, the mixture stirred for 1 hr., diluted with water (l ml.), and extracted with chloroform. The product, a dark yellow gum (19 mg.),had an IR spectrum similar to that of (21), with additional absorption at 1730

(acetate C=0) and 1550 cm,""**" TLC for this, and all other de amination products, was unhelpful, persistent tail ing being observed.

The acidic aqueous layer was poured on to cold saturated sodium hydroxide solution (4 ml.), and extracted with chloroform, yielding more of the above material (29 mg.). Similar products were obtained - 119-

after heating for 1 hr. after each addition of sodium nitrite,

or Dy leaving the reaction mixture stirring for 4 days before

extraction.

b) A solution of (21), (53 mg.), i*1 acetic acid (l ml.) and water

(2 ml.) was treated over 1 hr. with sodium nitrite (30 mg.), the

mixture stirred for 90 min., and warmed for 1 hr. on the steam bath.

Extraction as in (a) gave a very impure product similar to those

from (a).

1.5-dinitro,- 3«6-dimethyl~3-azabicyclo (3.3.l)nonan-7p-ol

To a suspension of sodium borohydride (JQ mg.; 15/8 mM

plus ca. 10$ excess) in dry TBF (5 nil.), and the aminoolefin (lb),

(1*205 gnu; 5 mM), was added dropwise over 30 min., with stirring

and in a nitrogen atmosphere, boron trifluoride etherate (357 mg.;

2*5 mM). The mixture was stirred under nitrogen for20hr. (2 hr.

stirring led to much lower yields), and water (2 ml.) was carefully

added, followed by saturated sodium hydroxide solution (3N; 1 ml.)

and aqueous hydrogen peroxide (3Q/>; 1 ml.). After being stirred

for 40 hr., the solution was extracted with chloroform, and the

organic extracts washed, dried, and evaporated, leaving a brown

crystalline solid (1*210 gra.). The product contained much starting material, together with the more polar alcohol (IR, TLC), separation being achieved by preparative TLC (thickness 0*6 mm,, eluant: light petroleum/30$ EtOAc, each plate run three times). The more polar -120-

band contained crystalline (26), (85 mg. from 720 mg. product; 11$), rn.p. 96-97*5° (colourless crystals from aqueous ethanol). 1R, table 2; RMR, table22.

Jones oxidation of (26)

The alcohol (26), (10 mg.) was treated in the usual way with excess Jones reagent, yielding a mixture of three components, all more polar than (26) < Routine IR showed much weaker hydroxyl absorption (3500 cm*~^), a carbonyl doublet at 1720 and 1680 cm.“^, and a nitro band at 1555 cm.“l l,5-dinitro-5-methyl-7-amido-5-azabicyclo ('5.3.1') non-6-ene (21)

The reaction mixture from 3»5~dinitrobenzamiae (2 gm.) was stirred for 2 hr, after acidification, suction-filtered, and the filtrate extracted with chloroform (4 X). The filtered solid was washed several times with water and dried, affording reasonably pure

(2$. (1.18 gm.-)rf The chloroform extract was washed, dried, and evaporated, leaving pale yellow crystalline material (482 mg.), largely insoluble in chloroform, identical to the filtered solid (IR)« The total yield was 1*702 gn. (67$)» m.p. 201-203° (colourless plates from ethanol).

Found: C, 44*53; H, 4*90; N, 20*62. C10H14N4O5 requires 0,44*45;

H, 5*22; N, 20*73*

IR, table 1; M®, table21; mass spectrum,table 12* 1,5~dinitro-3-methyl-3-azabic.yclo (3.5.1) nonan-7-one (23)

(a) from amide (2d)

To a stirred, solution of amide (3d), (3*24 gm.; 12mM),

in THF (35 ml*) and methanol (25 ml.), cooled by an ice-salt bath, 41,57 was added over 15 min. sodium hypochlorite solution (0*79QN;

18 ml„), the temperature being kept below -5°. The mixture was

allowed to warm to room temperature, stirred for 30 min., heated to

55-65° for 10 min., and dilute hydrochloric acid (6Nj 2 ml.) was

carefully added. The resultant clear red solution was refluxed for

30 min. on the steam bath, cooled, basified with solid potassium

carbonate, filtered, and extracted with chloroform (6 X). The

chloroform extract was washed and dried* Removal of solvents

afforded a partially solid red oil, which contained starting amide

(IR, TLC), together with five other compounds, two of which were

predominant (TLC). The amide was removed by several filtrations from

chloroform solution, followed by evaporation of solvent. This yielded

amide (203 mg.) and a clear red oil (2*688 gm.), further purified by

thorou^i extraction with hot carbon tetrachloride to a paler, more

mobile oil which partially solidified, The ketone (23) was isolated

by preparative TLC (thickness 1 mm., eluant: li^it petroleum /50$ EtOAc).

From the least polar band was obtained crude crystalline (23), (356 mg.;

12.3$ based on consumed amide), n.p, 120-121** (colourless needles from

carbon tetrachloride)• -122-

Foriid: 0,44*23; H, 5*45; Ify 17*22* C9H15NJO5 requires

c,44°45; H, 5-59; N, 17*28.

It formed neither a picrate nor a dibenzylidene derivative.

IR, table^,7; MR, table24*

(b) from vinyl ether (93) (See p. 129)

1.5-dinitro-3-nethyl-5-azabicyclo (3.3.1) nonan-7cC-ol (35)

A solution of aminoketone (23) , (73 mg.; 0*3 mM), in

ethanol (10 ml.) was treated portionwise with sodium borfihydride

(28 mg.), and stirred for 8 hr. Brine was then added, and the mixture extracted with chloroform. The organic layer was washed,

dried, and evaporated, leaving (33) as a colourless crystalline solid

(67 mg.; 91 %) t m. p. 159*5-162*5° (colourless needles from ethanol-

carbon tetrachloride).

Found: 0,43*91; H, 6*02; R,17*00. C9H15E3O5 requires

0,44*08; E, 6*17; N, 17*13*

From its IR and M R spectra (tables 2 and22respectively),

(33) was shown to have an end6 hydroxyl group.

1.5-dinitro-5-ethyl-5-azabicyolo(3.3.l) non-6-ene (54)

m-Dinitr^benzene (2 gm.) was reduced by sodium

borohydriie 111 the usual way (see preparation of (5b)). A

mixture of ethylamine (70‘fo aqueous solution; 5 ml.) in water (5 ml.),

formaldehyde (9 ml.) and water (9 ml.) was then added. After being -123-

stirred for 3 hr., the mixture was acidified with acotio acid

(8 ml.), stirred for a further 90 min., and extracted with chloroform;

The organic extract was washed (brine), and extracted with dilute hydrochloric acid, the acidic extract basified with solid potassium carbonate and extracted with chloroform. On removal of solvent, there remained a red-brown oil (1*18 gm<; 41i°) * Extraction with hot ethanol, cooling, filtration of precipitated black gum, re cooling, and filtration afforded (34) as a pale yellow crystalline sdid, m.p. 41-42° after further recrystallisation from ethanol.

In later runs, pure (34) extracted more easily by hot petrol

40-60°. Solubility in ethanol increased with purity, and the analytical sample was obtained by recrystallisation from light petroleum (60-80°, then 40-60°).

Pound: 0,49*78; H, 6*48; N, 17*60; C3.0H15N3O4 requires

0,49*79; H, 6*27; N, 17*42.

IR, table 4; MR, table35; mass spectrum, tablel3 *

Picrate, m.p. 153-159° (2 X ethanol/EtOAc).

1.5-dinitro-3-benzyl-5-azabicyclo (5« 3»l) non-6-ene (35)

The reduction of m-dinitrobenzene (2 gm.) by sodium borohydride, and addition of ice-water was carried out as in the preparation of (2b). Benzylamine (3 ml.) in water (6 ml.), formaldehyde (9 ml.) and water (9 ml.) were added together, and the -124-

mixture stirred for 2 hr. Acetic acid (9 ml.) was then added, and after being stirred for 1 hr., the solution was filtered and extracted with chloroform. The procedure used in the preparation of

(34) yielded a dark red oil (443 mg.; 12$), Attempted recrystallisations from a variety of solvents were unsuccessful. Pi crate, after re crystallisation from ethanol, had m.p* ca, 134-144° (some "sweating** commencing at ca. 125°).

Pound: C, 47*45; H, 4*04; R, 15*80. C21H20R6O11 requires

0 , 47*37; H, 3*79; », 15*78.

IR (of25), table 4 ; M R (of picrate), table 16« i,5-dinitr8-3-azabicyolO (5.3.1) non-6-ene (36)

Reduction of m-dinitrQbenzene (2 gm.) as in the preparation of (21), was followed by addition of a mixture containing concentrated ammonia (S.G* 0*88; 9 ml.), formaldehyde (9 ml*), and water (9 ml,), and stirring for 1 hr. On acidification with acetic acid (9 ml.), the colour of the solution changed from dark brown to orange, and a brown oil was precipitated. Stirring for 1 hr. (overni^it stirring gave a negligible yield of basic material), extraction with chloroform and isolation as for (34) furnished a dark yellow oil (1*01 gm.; 40$) which partially solidified. Routine IR showed NH absorption and TLC (using an amine buffer solvent, 14$ dimethylanine in benzene) showed the presence of two amines. The product was washed with hot ethanol, hot carbon tetrachloride, and cold ethanol (till the washings were colourless), -125-

leaving a straw-coloured solid, m.p* 132-140°, which could not b» recrystallised from carbon tetrachloride/chloroform or carbon tetrachloride/methanol. Picrate, m.p, ca, 150-185° (ethanol).

Further recrystallisation did not change the m.p, range, which appeared

to include twa melting points, TLC also indicated the presence of

two p orates, but insolubility in common solvents precluded the use

of preparative TLC,

IR, table 3; NMR, table17.

Attempted preparation of 1,5-dinitrO-5-phenyI-5-azabicyclO (3.3»l) non-6-ene (57)

Treatment of m-dinitrobenzene (2 gnu) as in the preparation of

(35), with aniline instead of benzylaraine, yielded a brown oil (340 mg,),

containing some aniline (odour,IR spectrum). The product was a

mixture of three compounds, the two main ones being less polar than Oa).

Attempted preparation of l,5-dinitro-3-acetyl-3-azabicyclo (3*3*1) non-6-ene (5§)

m-Linitrbbenzene (2 gm,) was reacted as in the preparation

of (36), with acetamide (l gm.) in place of ammonia. Chloroform

extraction afforded a red oil (740 mg.), which appeared to contain

two components, both more polar than starting material, and tailing

considerably (TLC). Its IR spectrum did not indicate the presence

of a tertiary amide, and no solid was obtained from a cooled ethanolic

extract. -126

lt5—dinitro-7-carbonethoxy-3-azabicyclo (3.3.1) non-6-ene

Methyl 3,5-dinitrobenzoate (2 gm.), treated as in the preparation of (36), gave a clear green solution on acidification with acetic acid. The chloroform extract was washed, dried, and evaporated.

The residue was a viscous pale yellow oil (1*85 gm. $ 77/0 which slowly crystallised. Washing with hot chloroform, then hot nethanol, gave (39) as a colourless powder, further purified by dissolving in hot chloroform, adding carbon tetrachloride and then light petroleum

40-60°, and cooling (using large volumes of solvents). The m.p. was 191*0-192.5o.

Found: C, 44*40; H, 4*66; N, 15*55. C10H13R3O6 requires

0, 44*28; H, 4*83? N, 15*49.

IR, table3 J MR, table21.

Hydrogenation of 1.5-dinitrO-7-carbonethoxy-3-azabioyclo (5.3.1) non-6-ene USE The aminoester (39), (73 mg,), dissolved in acetic acid (5 ml.), was hydrogenated for 7 hr. with Adams’ catalyst (40 mg,). Filtration and evaporation of solvent left a colourless viscous gum, too in soluble for TLC analysis. IR showed ester and nitro group absorption, and a peak at 1635 cm."1 replacing one at 1655 cm."*1 (C=C) in (39).

Attempted cyclisation of the hydrogenation product

A nujol mull of the hydrogenation product (of uncertain constitution) was heated for 20 min, at 180—200°, Some darkening was observed, and the resultant highly viscous material showed IR absorption at I64O and 1560 cm**"^-, the broad carbonyl peak having disappeared. Removal of nujol by washing with chloroform left a hard brown gum, the insolubility of which precluded TLC or further spectroscopic analysis* l,5-dinitro-7-amido-5-azabicyclo (5.5*l) non-6-ene

After treatment with the Mannich reagents used in the preparation of (36) > 3*5-

The solids were washed thoroughly with hot chloroform and hot methanol* giving a colourless powder, m.p, 234-235^» totally in soluble in all common solvents.

Found: C, 42*28; H, 4*6?i N, 22*02. requires

C, 42*19; H, 4*72; N, 21*87*

IR, table 3 ; MR, table21.

Attempted preparation df 1,5-dinitro-3-‘|5thyl-7~&mido-5-azabicyclo (5*3«1_) non-6-ene (A~5)

Following the procedure for the preparation of (34), 3»5~ -128-

dinitrobenzamide (2 gm.) yielded a viscous orange oil (ca. 100 mg,), after extracting the basified acidic extract with chloroform, followed by ethyl acetate. A crude red solid was obtained on filtering the reaction mixture prior to extraction. Although these products may have contained the desired amide (IR spectra), purification by recrystallisation was unsuccessful.

Attempted preparation of lt5-dinitro-3-benzyl-7-amidQ-3-azabicyoln ('5.3 .1 ) non-6-ene 15— ------—

Following the procedure for the preparation of (3£>), the reaction mixture from 3>5-

20 min. before and after acidification, was extracted with chloroform, then with ethyl acetate, to yield a yellow oil (l*15 gm.)• Although the product had the expected IR spectrum, TLC analysis showed it to be a mixture of several components* Recrystallisation from ethanol was unsuccessful.

1.5-dinitro-3-m6thyl-7-m£thQxy-3-azabicyclo (3.5.1) non-6-ene fe)

A solution of 3,5-dinitroanisole (2 gm.; prepared from 42 1.3.5-trinitrobenzene ) in THF (30 ml.) formamide (20 ml.) and methanol (10 ml.) was treated in the usual way (with cautious addition of acetic acid to control the violent effervescence in this case, followed by stirring for 30 min.)to yield (£e) as a red oil, solidifying on standing (l»70 6&*i &&io). Recrystallisation from ethanol was difficult (large wastage, also material sensitive to heat), -129-

and gave colourless crystals, m.p. 110-111'°, which slowly yellowed on exposure to light.

Found: 0,46*81$ H,5#79; N,16*39. G10H15N5O5 requires 0 ,46*69;

H ,5•88; N,16*33.

IR, table 1; MR, tablel5.

Hydrolysis of (2e) to the aminoketone (23)

The vinyl ether (2e), (182 mg.), dissolved in dilute hydrochloric acid (10 ml.) was heated for 2 hr, on the steam bath.

The cooled solution was basified with dilute aqueous sodium

carbonate and extracted with chloroform. The organic extracts were washed, dried, and evaporated to yield a crystalline solid

(172 mg.; IOO7Q identical with (23) prepared from the amide (2d).

I,5-dini tro-7-mo thoxy-3-azabic.yclo (3«5*l) non-6-ene (48)

Reduction of 3>5-dinitroanisole (2 gm.), dissolved in THF

(30 ml.) and ethanol (15 ml.),by sodium borohydride (2 gm.),was

followed by addition of ice-water (50 ml.) and 8 ml. each of ammonia,

formaldehyde, and water. After being stirred for 30 min., acetic

acid (8 ml.) was carefully added, with stirring for a further 30 min.

Filtration and extraction of the filtrate with chloroform gave a red

oil containing some solid material, which was removed by several filtrations and evaporations of the chloroform solution, This material was not the desired product (IR spectrum). The red oil was acid- washed in the usual way to extract basic material (720 mg.) which -130-

slowly solidified* Routine IR showed the presence of NH, C=C and

NO2 groups, and also RCH3. It consisted of two compounds, the less polar predominating. Separation was accomplished by preparative

TLC (thickness Irani#, eluant: light petroleum /307° EtOAc). The main (less polar) band yielded 363 rag* crystalline solid, identical to the N-nethyl analogue (2^ of the desired product (IR, TLC, m*p» after recrystallisation).- From the other band was obtained

(4.8) a pale yellow oil (160 mg,; 6* 57-0» which did not solidify.

IR, table 3; MR, table23.

1.5-dinitro-3-ethyl-7-methoxy -3-azabicyclo (3«5»l)non-6--ene (49)

3,5-dinitroanisole (2 gm.) when treated as in the preparation of (34), afforded a red oil (l«79 gm*), which consisted of one predominant compound, and several more polar ones (TLC),

Ry means of preparative TLC (thickness lmn., eluant: light petroleum

/ZQffo EtOAc, each plate run twice), the major constituent (49) was obtained as a pale yellow oil (283 mg.; 10*370*

IR, table4? HMR, table23» Picrate, m.p, ca. 143-158° (3 X EtCH).

1.5-dinitro-3-benzyl-7-methoxy -5-azabicyolo (3.3.1) non-6-ene (5_Q)_

Following the procedure used for the preparation of (35),

3.5-dini troanis ole (1 gm.) formed a red oil (l*315 8ih), consisting of two compounds, the less polar one predominating. Separation by preparative TLC (as for (49)) gave (50) as a pale yellow oil

(407 mg.; 24^). -131-

IR, table 4 ; IMR, table23, Pi orate, m.p, ca. I40-I520 (2 X EtCK).

1 .5-dinitro-5-azabicyelo (3,3*1) nonan-7-one (45)

The dinitrovinyl ether (4 9 . (100 mg,), was hydrolysed as in the preparation of (2 3), forming a yellow oil (73 mg,; 70^), which crystallised on cooling. M.p. range' was wide, final n,p. being ca. 170°, Recry a tall is at io n was however unsuccessful.

IR» table 6,7®

1.5-dinitrc-3-ethy1-3-azabioyclo (5«3.l) nonan-7-one (46)

Hydrolysis (as for (48)) of (49), (100 mg.), afforded (46) as a colourless crystalline solid (71 mg.; 7570* m.p. 121-123°

(colourless needles from carbon tetrachloride).

Found: 0,46*71; H, 5*83; N, 16*31* C1O015N3O5 requires

C, 46*69; H, 5*88; N, 16*33*

IR, table6,7? MR, table24*

1.5-dinitro-3-benzyl-5-aznbioycio (3*3*1) nonan-7-one (47)

The product (47) from hydrolysis (as for ( £ ) ) of (50),

(100 mg.) was a solid, which precipitated from the aqueous acidic solution on cooling, and was totally extractable without ba.sifi.cation,

(6? rag.; 70/0 . m.p* 155-157° (colourless needles from carbon tetrachloride).

Found: 0,56*65; H,4*81; 1,13*26. requires

0,56*42; H,5*37* »,13*16. -132-

IR, table^ MR, table24,

Attempted preparation of 1,5-dinitro-2,4-diphenvl-3-nethvl-5- azabioyolo (3.3.1) non-6-ene

in-Dinitrobenzene (2 gm.) was reduced in the usual way.

Benzaldehyde (2*5 gm.), then methylamine (33$ in EtCH; 8 ml.), was added with cooling. ,ifter being stirred for 42 hr., the opaque red solution was acidified with acetic acid (8 ml.). Extraction of basic material gave a brown oil (680 mg.), a complex mixture of ca. six components (TLC).

Attempted preparation of 1,5-dinitro-2,3»4-trinethyl-3-azabicyclo (3.3.1) non-6-ene

m-Binitrobenzene (2 gn.), after reduction, was treated with a mixture of acetaldehyde (4 ml.) in water (4 ml.), methylamine (33$ aqueous;

8 ml.) and water (8 ml.). After being stirred for 2 hr., the mixture was acidified with acetic acid (8 ml.), stirred for 1 hr., and filtered,

Basic material was extracted as a viscous red-black gum (162 mg»), consisting mainly of one impure compound (TLC). It could not be obtained solid, nor could a picrate be prepared.

Attempted preparation of the azatricyolododecane (511

Reduction of m-dinitrobenzene (2 gm.) with sodium borohydride

(1*5 gm.) was followed by treatment with a mixture of glutaraldehyde

(25$ aqueous solution; 10 ml.), methylamine (10 ml.), and wat^r (10 ml.).

The mixture was stirred for 1 hr., acidified with acetic acid (10 ml.), -133-

and stirred for a further 1 hr« Extraction of basic material yielded a brown oil (22 mg»), which was not further investigated.

Attempted preparation of the azatricyclododecane (52)

Under similar conditions, methyl 3,5-dinitrobenzoate (2 gm.)

formed a brown gum (115 mg.). Routine IE showed the presence of

ester and nitro groups, but TLC indicated a complex mixture (much of

the material remaining on the baseline).

Attempted condensationsbetween the disodium salt (53) and electrpphiles, aiming at 1 ,5-dinitrobicyclo (5.5.1) non-2-enes

a) ACROLEIN ------33 To a suspension of the salt (33) in the solvent mixture was

added dropwise at 0° acrolein (equimolar amount), with or without the

addition of base. The mixture was stirred for several (usually

2-4) hours, acidified with acetic acid, and extracted with chloroform.

In all cases, the products (red or black oils) had peaks

in the IR due to 0=0, NOg and sometimes OH, indicating that some

condensation had occurred. Large amounts bf starting material were

recovered in the absence of base, but in the presence of hydroxide or

ethaxide, these amounts were greatly decreased. Also present were

two or three other components (TLC), e.g., after addition of 1:1

acrolein: ethoxide, and stirring for 65 hr., no starting material was

present, but the crude product was a complex mixture. No purer product was obtained by isolating the solid salt (53), (under the last- -134-

mentioned conditions), stirring for 13 hr*

Distillation of the product was ineffective, due to charring. b) A CRYLONITRILE

Addition of acrylonitrile (under conditions described for acrolein) yielded after ether extraction a dark red oil. Routine

IR showed nitrile, nitro, and carbonyl absorption, and TLC analysis indicated a mixture of three components. Distillation afforded with large loss an orange oil, b.p, ca. 30o/Ch03 nsft., with nitrile, but no nitro absorption in the IR, i.e. decomposition had occurred, c) PHENYL VINYL KETONE (FVK)

The salt (53) was filtered and washed with dry ethanol.

A solution of PVK (equimolar amount; obtained by distillation of the Mannich base hydrochloride) in dry ether was added dropwise over 10 min. under nitrogen to a stirred suspension of (53) in dry ethanol at -70°. The mixture was stirred for 1 hr., aliquots being removed after 5»20, and 40 min. After addition to ice and acidification with dilute tartaric acid, extraction with chloroform gave products with identical highly complex thin layer chromatograms,

IR contained peaks at 1680 (broad; C=C and/ or FhCO), 1600, 1580

(ar. C=C) and 1550 (NO2) cm."1 Unsuccessful efforts were made to obtain solid products, 1-methyl-2.4-dinitrocyclohex-l-ene 33 fed I 2 ,4**dinitrotoluene (4 gm.) reduced by sodium borohydride -135-

(4 gm.) and the mixture acidified with 1C% tartaric acid, afforded a mobile red oil (3*45 gm.). Routine IR showed a weak peak at

1680 (C=C), and a broad one at 1560 cm* ^ (NC^). The corresponding peaks in the starting material were at 1605 and 1540 cm. ^ TLC analysis however indicated the presence of three components (two compounds predominating, possibly double bond isomers). The product became lighter in colour on distillation (b.p. ca, ll6°/o«l mm.; Lit.

75°/o*i mm,, then analysed), the yield being 1*90 gm. (Lit. yield not quoted). IR then showed a strong peak at 1685 cm.~^ (C=C), but the material was of unchanged composition (TLC). After several days, the IR peak at 1685 cm."”^ was of much weaker intensity^ ite*, the material was unstable and reverting to its undistilled state.

Attempted preparation of l-carboxy-3»5-dinitrocyolohex-l-ene (56) 33 Under the conditions described in the reduction of the

2 ,4-dinitro acid, 3>5-

Attempted preparation of 1-carbethoxy—3>5-dinitrocyclohex-l—ene (57)_

Ethyl 3 ,5-dinitrobenzoate (2 gm.) was reduced as in the - 1 3 6 -

prep? ration of (2b). Dilution with ice-water (200 ml.), acidification with 10$ tartaric acid (50 ml.) and stirring for 12 hr., was followed by extraction with methylene chloride. The organic extracts were washed with saturated sodium bicarbonate solution, leaving neutral material as a dark red oil. Routine IRs 1725 > 1550» 750 cm.~^

(starting material 1725, 1630, 1550, 730 cm.**-*-). TLC analysis

showed the product to be a mixture of several compounds, considerable

streaking occurring on chromatograms. 13,20 l-cerbethoxy-5-methyl-3-a2abicyclo (5*3»l) nonan-9-one (3 8 )

2-carbeth.oxy cyclohexanone (6*0 gm.; 0*035 M) afforded a

basic pale brown oil (4*6 gm.; Lit. 7*2 gm*)> which had the expected

IR spectrum, but consisted of about four components (TLC),

Treatment of the product with ethanolic picric acid gave a brown

gumi Purification was accomplished by preparative TLC (thickness

1 mm,, eluanti light petroleum/25$ EtOAc). Extraction of the

first main layer above the baseline yielded reasonably pure (58)

(one main spot TLC) t The recovery was 25-35$> representing a yield

of 15-21$. Mass spectrum (P*225 n/e) showed peaks at 224 (P-l)>

196 (P-29, loss of C2H5), and 180 (P-45? loss of OEt) m/e.

IR, table 5; HMR, table25*

Attempted preparation of l-carbethoxy-3—nethyl-3—azabicyclo (3*2.1). octan-8-one (6 6 j

Using the procedure for the preparation of (58), - 137 -

2-carbethoxycyclopentan©ne (5*46 gm,; 0*035 M) formed a pale-brown basic fraction (l«17 g&*). TLC analysis showed four compounds to be present, the two main ones having Rf values similar to (58).

Separation by preparative TLC (as for (58)), and examination of the

IR spectra of the two compounds, showed however that neither was the desired compound (66)* l-hydroxymethyl-5-methyl-3-azabicyclo (3»5 .l) nonan-9-ol (61 )

A suspension of lithium aluminium hydride (1*0 gm.; 30 mM;

3M excess) in dry THF (15 ml.) was added carefully dropwise to a solution of ketoester (58), (2*25 gra»> 10 mM), in dry THF (35 ral«)> and the mixture refluxed for 15 hr. Remaining LAH was destroyed by cautiously adding moist ethanol* A large volume of brine wag added, and the mixture extracted with chloroform. The organic extracts were washed, dried, and evaporated to leave (61) as a viscous pale yellow oil (1*52 gm.; 82$). A sample was distilled to yield a viscous colourless liquid* b.p. ca. 125-130°/lmm,, which slowly solidified. Several recrystallisations from light petroleum 60-80°, then ligit petroleum 60-80°/carbon tetrachloride, afforded a colourless solid, m.p* 105—131° (C9 epimers)• Further recrystallisation did not raise the m.p., and satisfactory analytical figures were not obtained. The mass spectrum had a parent ion at P=185 n/e.

IR, table 8 ; NMR» table26. -138-

1-tosyloxymethyl-6-nethylbicvclo (4,5.l)-dec-7-en-10-ol (71) 61 The cliol (68), (980 rag.; 5mM), dissolved in dry pyridine

(4 si.) was treated with a solution of tosyl chloride (970 ng.; 5nM plus slight excess) in dry pyridine (4 ail.). The mixture was kept at 0° for 3 days, and extracted aS in the preparation of (43)^ except for s idification with dilute hydrochloric acid prior to chloroform extraction. The product, a yellow oil (1728 mg.; 99$) partially solidified, and consisted (TLC) of two compounds (CpQ epimers).

Separation by preparative TLC (thickness 0*75 nm., eluant: light petroleum/25$ EtOAc) gave the more polar epimer as a yellow oil (4O8 mg.), and the less polar epiner as a yellow oil (1083 ng.) which solidified, m.p. 130-1370 (light petroleum 60-80°/carbon tetrachloride).

IR, table 9; MR, table

62 l-tosyloxymethyl-5~meth.yrbicyclo (3.3.1) non-3-en-9-ol (72 ) 61,62 The diol (69), (prepared by reduction of the corresponding ketoester; 98$, Lit, 92$) (910 ng.; 5 mM)> under the conditions described in the preparation of (71) > afforded a ser.ii-solid pale yellow oil, containing both C9 epimers (TLC). Preparative TLC (thickness

0*75 nn., eluant: light petroleun/l5$ EtOAc, each plate run twice), yielded the less polar epiner as a crystalline solid (817 ng.), m.p.

71-72° (light petroleum 60-80°).

Pound: C, 64*77; H, 7*44. C18K2404S squires C, 64*26; H, 7*19- The more polar component (140 mg.), which did not solidify, contained

a trace of its epiner (TLC). The yield of pure (72) was 64$

(Lit. epimeric mixture had m*p. 65-67°* yield 90$)..

IR, table 9; MR, table27.

1-tosyloxynethy 1-6-methylbicvclo (4.3.1) decan-10-ol (73) 61 The diol (70), (434 mg.), formed on tos.ylation a pale yellow seni-solid oil, which when washed with light petroleum

60-80° became a colourless crystalline solid (440 ng.), n.p. 106*5°”

108° (light petroleum 60-80°), This solid appeared to be a single

compound (TLC),

Pound? 0,64*88; H, 7*87. C19H28O4S requires C, 64*75? H, 8*00.

The washings contained a yellow oil (240 ng.), consisting of (73)

and a trace of diol (TLC). The yield was } 73$ (assuming *>nly half

of petrol-soluble material to be (73))*

IR, table 9; MR, tabled.

1-10syloxymethyl-3-methyl-5-azabic.vclo (3.3.1) nonan-9-ol .(^?).

A distilled sample of the diol (6l)» (370 mg.; 2 mM),

dissolved in dry pyridine (l ml.) was treated with a solution of

tosyl chloride (390 mg,; insLight excess of 2 mM) in dry pyridine.

After being left at 0° for 3 days, the mixture was added to ice,

and extracted with chloroform. The chloroform layer was washed,

dried, and evaporated to yield a brown oil (279 mg.). TLC analysis -140-

indicated the presence of a trace of diol (even after the reaction was left for 7 days at 0°), a less polar compound (62), and a least polar compound* Separation by preparative TLC (thickness 1 mm., eluant: EtOAc) afforded (62) as a brown oil (68 mg*; 10$) which did not solidify. IK* table 9? MR, table 27. l-tosyloxymethyl-6-nethylbicyclo (4.3.1) dec-7-en-10-one (74)

The less polar epimer of (71), (393 mg.), on treatment with excess Jones reagent afforded a yellow oil (337 mg.), consisting of two compounds with very similar Rf values (TLC), Separation was achieved by preparative TLC (thickness 0*75 mm., eluant: light petroleum /7*5$ EtOAc, each plate run six times), affording the less polar material (74) as a crystalline solid (224 mg.; 57$ ) t definitely only one compound (TLC), with m.p, 82*5-83*5° (light petroleum 60-80°).

Pound: C, 65*22; H, 6*91. C19H24O4S requires C, 65*50; H, 6*94.

IR, table 10; MR, table28.

The more polar component (41 mg.), also a crystalline solid, possessed the CH20Ts group (IE, MR) but lacked the O C (MR) and C=0 (IR) groups. It was not further investigated.

Under similar conditions, the more polar epimer of (71) formed a product that contained ca. 20$ acidic material. The neutral product was a complex mixture (TLC)• These products were not further examined. -141-

1- t.ogyloxymethyl-5-methylbicvclo (3.3*1) non—3— o*»-9—one (75)

Both epimers of (72), when oxidised by Jones reagent, yielded identical products (IR, TLC, m*p.) (yields: more polar epimer 89$, less polar, 93$). M.p* 65-68° (light petroleum 60-80°).

TLC chromatograms, run five tines in light petroleum /lO$ EtOAc,

showed that two compounds of almost identical Rf were present. These might be A double bond isomers, formed by the acidic oxidation

conditions (this was not however observed in the case of (74)).

Pound: C, 64*60; H, 6• 63* 8]_8^22^4 ® requires C, 64*65; H, 6*63.

IR, table 10 MR, table28.

l-tosyloxymethyl-6-methylbicyclo (4.3.1) decan-10-one (76)

The nonotosylate (73), (23 mg.), when treated with excess

Jones reagent, furnished a colourless oil (227 mg.; 99$)* A sample

separated from trace impurities by preparative TLC (thickness 0*25 mm., eluant: li^it petroleum / lG $ EtOAc) solidified, and had m.p. 74-76°

(ligiht petroleum 60-80°). Mass spectrum (P=350 m/e) had Pea^s at

195 (loss of T$, 155), 179 (loss of OTs, 171) and 178 (loss of TsGH,

172) n/e.

IR, tablelO; MR, table2&

1—tosyloxymethyl—3—methyl-3-azabicyclo (3»3»l) nonan-9-one (63)..

A solution of the nonotosylate (62), (380 mg.) in acetone

(10 ml.) was treated with excess Jones reagent with stirring at 0°. - 142 -

After 30 min., excess methanol was added, then excess brine, and the mixture extracted with chloroform. The organic layer was washed, dried, and the solvents removed, leaving a brown oil (288 mg.).

Neutralisation of the acidic aqueous layer with saturated aqueous sodium carbonate, followed by chloroform extraction, yielded further product (21 mg.). The total yield was 304 mg. (81$). Samples for spectral data were freed from more polar impurities by preparative

TLC (thickness 0*25 mm., eluant: light petroleum/l5$ EtOAc) which furnished (63) as a brown gum which did not solidify.

IR, tablelO; NMR, table 28.

Treatment of the ketotosylates (74). (75) and (76) with strong base a) Sodium ethoxide

A solution of the ketotosylate in dry ethanol was stirred under reflux with a solution of sodium ethoxide in dry ethanol.

Water was added to the cooled solution, and the acidified (dilute

HCl) mixture was extrac ted with ether. The organic extract was washed with saturated aqueous sodium bicarbonate, brine, and dried.

The ketotosylate (76), (150 mg.) in dry ethanol (9 ml.) was

treated 16 hr. as above with sodium (15 mg.) in dry ethanol (5 ml.)*

The product (90 mg.) contained unchanged (76), a more polar component

(predominating), and a less polar component. IR showed strong hydroxyl

and carbonyl absorption at 3450 and 1710, 1695 cm.”1 respectively.

More prolonged treatment (50 hr.) afforded similar material. Separation by preparative TLC (thickness 0*25 mm., eluant j light petroleum

/$0$ EtOAc) yielded the most polar constituent (25 mg.) as a colourless oil. IR (CCI4): 3630 (free OH), 3515 broad (bended OH), doublet at 1697, 1686 (0=0). The evidence indicated structure

(77)* Treatment of this material with tosyl chloride in pyridine furr shed a product containing unchanged (77 ) and a less polar compound with a Rp value similar to that of (76),

The ketotosylate (75) when treated as above for 50 hr., remained almost totally unchanged. When refluxed for 4 days, however, the product consisted of starting material, one less polar, and two more polar compounds. Here again, IR showed strong hydroxyl and carbonyl absorption at 3500 and 1710 cm.“^, respectively, as well as absorption due to (75)* b) Potassium t-butoxide

The procedure used was similar to that in (a), with potassium in t-butanol instead of sodium in ethanol.

The ketotosylate (76) after 19 hr. afforded a mixture of compounds, containing no (76), but three less polar components and one more polar. Since the IR spectrum (3500s, hydroxyl, and 1700s, carbonyl) was similar to that of (7 7 ), it appeared that the most polar constituent was the ketol (77). Similar IR absorption was exhibited by the product from (74) -144-

after 14 hr. reflux. This consisted of ca. three compounds, two

heing nore polar than (74) , the less polar of these predominating.

By analogy with (76) > the latter was probably the ketol (79).

The product after 17 hr. from (75) contained equal amounts

of acidic (probably two acids) and neutral material. The latter was a

highly complex mixture, with IR absorption similar to that quoted

above• -145-

REFEHENCES

1. Wiesner and Valenta, Progr. Chem. Org. Hat. Prod., 1958, 26.

2. Komppa, Ber., 1952, 6£, 792.

3* Rossi and Valvo, Farmaco (Pavia), 1957, 12, 1008; Chem. Abstr., 1958, £2, 12860i.

4* Rossi, Butta, and Valvo, ibid., 1959> 14> 666; Chem. Abstr., i960, 6710e.

5. Rossi, Valvo and Butta, Gazz. Chim. Ital., 1959, 8£» II64.

6. Rossi and Valvo, Chim. Ind. (Milan), i960, £2, 6375 Chem. Abstr., 1964, 60, 11984c.

7* Rice and Grogan, J. Org. Chem., 1958, 2j5, 844.

8, Schneider and Gotz, Naturwiss., i960, 61.

9. Schneider and Gotz, Arch. Pharm., 1961, 294. 506.

10. Leonard, Conrow and Sauers, J. Amer. Chem. Soc., 1958, 80, 5185*

11. Mannich and Muck, Ber., 1930, 65. 604.

12. Mannich and Mohs, ibid., 1930, 63. 608; Mannich and Veit, ibid., 1935» 88, 506.

13. Weatherbee, Adcock, and Winter, J. Org. Chem., 1957» 22, 465*

14. Blicke and McCarty, ibid., 1959> 2£, 1379*

15. Blicke and McCarty, ibid., 1959> 1°69*

16. Schneider and Gotz, Naturwiss., I960, ££, 397.

17. House, Wickham, and Mueller, J. Auer. Chem. Soc., 1962, Q±, 3139.

18. Rossi and Butta, Ann. Chim. (Rome), 1962, £2, 381; Chem. Abstr., 1962, 9810i.

19. Iwai, Ogiso and Shimizu, Chem. and Ind., 1962, 1288. -146-

20* Shimizu, Ogiso and Iwai, Chem. Pharm* Bull. (Tokyo), 1963,11, 3

21* Shimizu, Ogiso and Iwai, ibid., 1963*11, 766, 770, 774*

22* Becker, J. Prakt, Chem., 1964* 2£, 259*

23. Bohlinann, Ber., 1958, 91, 2157*

24* House and Mueller, J. Org. Chem., 1962, 2^, 4436.

25* House, Mueller, Pitt and Wickham, ibid., 1963* 28, 2407*

26. House and Bryant, ibid., 1965, 30, 3634*

27* House and Pitt, ibid., 1966, j51, 1062; House and Tefertiller, ibid., 1966, J51, 1068; House, Tefertiller and Pitt, ibid., 1966, jHL, 3120.

28. House and Bryant, ibid., 1966, Jl, 3482.

29. Pumphrey and Rooinson, Chem. and Ind., 1963, 1903*

30. Dobler and Dunitz, Helv. Chim. Acta, 1964, 695*

31. Lygo, McKenna, and Sutherland, Chem. Connun., 1965, 356.

32. Severin and Schmitz, Ber., 1962, 1417*

33* Severin and Adam, ibid., 1963, .26, 448.

34* Severin, Schmitz and Temae; ibid., 1963, 5J6, 2499*

35* Severin, Schmitz and Adam, ibid., 1963, 26» 3076.

36. Severin and Schmitz, ibid., 1963, ^6, 3081; Severin and Adam, ibid., 1964, J9Z» 186.

37. Severin, Schmitz and Tenure, ibid., 1964, .92, 467? Severin and Temme, ibid, 1965, 98, 1159*

38. Borsche, ibid., 1923, ^6, 1488.

39. Buehler and Calfee, Ind. Eng. Chem. Anal., 1934, £» 351*

40. Chapas, Bull. Soc. Chim. Fr., 1927, 193* - 147-

41. We-erman, Ann., 1913, 401, 1.

42. Org. Syntheses, Coll* Vol. I, 219,

43. Fewer and Anderson, J* Amer. Chem. Soc., 1961, 8jj[, 2960.

44* Bryant, Burlingame, House, Pitt, and Tefertiller, J. Org* Chem,, I966, j>l9 3120; Duffield, Budzikiewicz, Williams, and Djerassij J, Amer. Chem. Soc,, 1965* 82, 810*

45* Buchanan and Curran, unpublished work,

46, Ma and Warnhoff, Canad. J. Chem., 1965, 42* 1849? Kawazoe and Ohnishi, Chem. Pharm. Bull., 1963, 11, 243*

47, Nielsen, J. Org. Chem., 1962, 22, 2001.

48, Rabjohn, Org. Reactions, 331.

49* Scott,Interpretation of the Ultraviolet Spectra of Natural Products, Pergamon Press, 1964, pp. 77-78* 50. Baran, J. Org. Caem., i960, 2^, 257*

51. Ireland and Mander, Tetrahedron Letters, 1965, 2627*

52. Hassner and Heathcock, J. Org. Chem., 1964, 22, 1350, and references cited therein.

53* Gaylord, Pieduction with Complex Metal Hydrides, Interscience, 1956, p.p. 762-773.

54* Berson and Ben-Efraim, J. Amer, Chem. Soc., 1959, 81, 4094*

55* Stetter and Goebel, Ber., 1963, 96, 550.

56. Zweifel and Brown, Org. Reactions, ii. 1; Brown, Tetrahedron, 1961, 12, 117.

57* Bell and Archer, J. Amer. Chem, Soc., I960, 82, 4642.

58. Leonard, Fox, and Oki, ibid., 1954, 2£» 5708.

59. cf, M.A. Eakin, Ph.D. Thesis, Glasgow University, 1967*

60. cf, Garblsch, J. Amer. Chem. Soc., 1963, 85* 1696. - 148-

61. Buchanan, McKillop, and Raphael, J. Chem. Soc., 1965? 833.

62. Martin, Parker, and Raphael, ibid., 1964, 289*

63. Anet, Hughes, Marmion, and Ritchie, Austral. J. Sci. Res.,

1950, 2£, 330. SUTM ART $~~. THZS\$~~

~~Part I : Synthetic approaches to allohimachalol~~

2-(&-fornyibutyl)-2,6 f 6-trimethylcycloheptanone was synthesised in six steps from tetrahydroeucarvone* Attempts to effect aldol cyclisation to 1,5,5,8-tetramethylbicyclo (4.4*1) undec-7-en-ll-one, which has already been related to the naturally occurring sesquiterpene alcohol allohimachalol, were unsuccessful.

~~Other approaches afforded a variety of compounds, in particular, a number of bicyclO (4 .3*1) decanes derived from 1,5,5-trimethyl-~~

~~8-oxabicyclo (5*4*0) undec-6-en-9-one.~~

~~Part II x Studies in the 3-£,zabicyclo (3*3«l)nonane system~~

The reported procedure for the syntheses of a small number of 6-substituted l,5-dinitro-3-methyI-3-azabicyclo (3.3»1)non-6-enes was extended to provide a wide range of 6- and 7-substituted analogues, the structures of which have been rigorously proved by chemical and spectroscopic methods.

Initial difficulty in assigning certain M R signals of these compounds led to a study of the M R shifts of protons oc and JB to the nitrogen atom in simple amines and amine pi crates. Some correlation between structure and magnitude of shift m s observed.

The scope of the reaction sequence leading to these compounds was thorou^ily investigated, and a number of 3-unsubetituted, 3-ethyl, and 3-benzyl analogues prepared. In particular, this made available several 3-substituted l^-dinitro-^-^zabicyclo (3*3*l)nonon-7-ones,

~~IR studies of which established the absence of interaction between the amine and carbonyl functions.~~

~~Evidence was presented confirming reported findings of the chair-chair conformation of the 3-&zabicyclo (3*3»l)nonane system.~~